key: cord-0955777-lt17qxq0
authors: Pei, Rongjuan; Feng, Jianqi; Zhang, Yecheng; Sun, Hao; Li, Lian; Yang, Xuejie; He, Jiangping; Xiao, Shuqi; Xiong, Jin; Lin, Ying; Wen, Kun; Zhou, Hongwei; Chen, Jiekai; Rong, Zhili; Chen, Xinwen
title: Human Embryonic Stem Cell-derived Lung Organoids: a Model for SARS-CoV-2 Infection and Drug Test
date: 2020-08-10
journal: bioRxiv
DOI: 10.1101/2020.08.10.244350
sha: bc8e62cf1bdd3f883493796d8ccfdce26f57fa6d
doc_id: 955777
cord_uid: lt17qxq0

The coronavirus disease 2019 (COVID-19) pandemic is caused by infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which is spread primary via respiratory droplets and infects the lungs. Currently widely used cell lines and animals are unable to accurately mimic human physiological conditions because of the abnormal status of cell lines (transformed or cancer cells) and species differences between animals and humans. Organoids are stem cell-derived self-organized three-dimensional culture in vitro and model the physiological conditions of natural organs. Here we demonstrated that SARS-CoV-2 infected and extensively replicated in human embryonic stem cells (hESCs)-derived lung organoids, including airway and alveolar organoids. Ciliated cells, alveolar type 2 (AT2) cells and rare club cells were virus target cells. Electron microscopy captured typical replication, assembly and release ultrastructures and revealed the presence of viruses within lamellar bodies in AT2 cells. Virus infection induced more severe cell death in alveolar organoids than in airway organoids. Additionally, RNA-seq revealed early cell response to SARS-CoV-2 infection and an unexpected downregulation of ACE2 mRNA. Further, compared to the transmembrane protease, serine 2 (TMPRSS2) inhibitor camostat, the nucleotide analog prodrug Remdesivir potently inhibited SARS-CoV-2 replication in lung organoids. Therefore, human lung organoids can serve as a pathophysiological model for SARS-CoV-2 infection and drug discovery.

The current fast-evolving coronavirus disease 2019 pandemic is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which infects lungs and can lead to severe lung injury, multiorgan failure, and death [1] [2] [3] . To prevent and effectively manage COVID-19, public health, clinical interventions, basic research, and clinical investigation are all emergently required. For basic research, it is essential to establish models that can faithfully reproduce the viral life cycle and mimic the pathology of COVID- 19 .

Cell lines and animals are two major models for coronavirus infection in vitro and in vivo, respectively [4] [5] [6] [7] . Cell lines can be used to amplify and isolate viruses (like Vero and Vero E6 cells 8, 9 ) , to investigate the viral infection (like primary human airway epithelial cells, Caco-2 and Calu-3 cells 3,5,10,11 ), and to evaluate therapeutic molecules (like Huh7 and Vero E6 cells 12 ). Animal models can be used to mimic tissue-specific and systemic virus-host interaction and reveal the complex pathophysiology of coronaviruses-induced diseases 7 . Mice, hamster, ferrets, cats, and non-human primates have been reported to model COVID-19 [13] [14] [15] [16] [17] [18] [19] [20] [21] . These cell and animal models have greatly enriched our understanding of coronaviruses and assisted in the development of a variety of potential therapeutic drugs 7 . However, these models yet have obvious limitations. Species differences make animal model results unable to be effectively translated into clinical applications 22, 23 . Species differences (cells from species other than humans, like Vero cells) and abnormal status (transformed or cancer cells) make cell models unable to faithfully reproduce the viral infection cycle and host response 24-26 .

Organoids are a three-dimensional structure formed by self-assembly of stem cells in vitro 27, 28 . As the cell composition, tissue organization, physiological characteristics, and even functions are similar to natural organs in the body, organoids have been used for human virus studies 29, 30 . For SARS-CoV-2, lung, kidney, liver, intestine, and blood vessel organoids have been reported to be sensitive for virus infection [31] [32] [33] [34] [35] [36] [37] . Here using human embryonic stem cells (hESCs)-derived lung airway and alveolar organoids, we demonstrate that SARS-CoV-2 infects ciliated cells, alveolar type 2 cells (AT2 cells) as well as rare club cells, and remdesivir is more potent than camosat to inhibit virus infection.

Based on our previous protocol 38 , as well as other reported protocols 39,40 , we developed an optimized method to differentiate human airway organoids (hAWOs) and alveolar organoids (hALOs) from hESCs, which contained six stages, embryonic stem cells (ESCs), definitive endoderm (DE), anterior foregut endoderm (AFE), ventralized anterior foregut endoderm (VAFE), lung progenitors (LPs), and hAWOs and hALOs (Fig. 1a, b) . Quantitative RT-PCR revealed the expression dynamics of marker genes along differentiation ( proliferating cells (Ki67 + ) (Fig. 1e) . And hALOs contained AT2 cells (SPC + ) and AT1 cells (PDPN + or AQP5 + ) (Fig. 1f ). Since ACE2 is the receptor for SARS-CoV-2 for host cell entry and TMPRSS2 is the serine protease for spike (S) protein priming 5,9 , we checked their expression along the differentiation and found they were highly expressed in hAWOs and hALOs (Fig. 1g) .

To test whether SARS-CoV-2 infects human lung organoids, hAWOs and hALOs (ranging from day 31 (D13) to D41) were exposed to SRAS-CoV-2 at a multiplicity of infection (MOI) of 1. Samples were harvested at indicated time points after infection and processed for the various analyses shown in Live virus titration on Vero E6 cells and quantitative RT-PCR of viral RNA in the culture supernatant and cell lysates showed that hAWOs and hALOs were productively infected by SARS-CoV-2 ( Fig. 2a, b) . Viral RNA and infectious virus particles could be detected as early as 24 hours post infection (hpi), increased at 48 hpi, and remained stable at 72 hpi. Compared to hALOs, hAWOs produced less virus at 24 hpi and similar amount of virus at 48 hpi and 72 hpi (Fig. 2a, b) . Co-immunostaining of viral nucleocapsid protein (NP) and pan epithelial marker E-CAD showed that SARS-CoV-2 infected epithelial cells in human lung organoids (Fig. 2c) . Quantification analysis showed that the percentages of infected hAWOs increased from about 50% at 24 hpi to about 75% at 72 hpi (Fig. 2d) . And the percentages of infected cells within a single hAWO increased from about 24.9±3.7% at 24 hpi to 63.9±6.1% at 72 hpi ( Fig.   2e ). For hALOs, the percentages of infected organoids remained stable at about 85% and the percentages of infected cells per organoid remained about 30%-40% from 24 hpi to 72 hpi. These cellular infection results were consistent with viral RNA detection and infectious viral particle titration results.

To determine the cell tropism of SARS-CoV-2, we co-stained each cell lineage marker with viral N protein and virus receptor ACE2. Microscopy analyses revealed that ciliated cells (a-TUB + ) and alveolar type 2 cells (Pro-SPC + ) were the major target cells (Fig.3a, b, and Fig. S1 ), which was consistent with the previous report 41 . In addition, rare club cells (CC10 + ) could be infected ( Fig.3a) . In hAWOs, about 90%-95% infected cells were ciliated cells and about 5%-10% were club cells, and no basal (P63 + ) or goblet cells (MUC5AC + ) were found infected (Fig. 3c) . In hALOs, 100% infected cells were AT2 cells and no AT1 cells (PDPN + ) were found infected (Fig. 3c) . We also measured the percentages of infected cells within ciliated cells and AT2 cells. About 26±3.6% at 24 hpi and 64.5±9.8% at 72 hpi of ciliated cells were infected, and the percentages of infected AT2 cells remained stable at about 30%-40% from 24 hpi to 72 hpi (Fig. 3d, e) . The distinct infection dynamics of ciliated cells and AT2 cells indicated that more and more ciliated cells could be infected by SARS-CoV-2 during a prolonged infection period and even all the ciliated cells could be finally infected when given long enough infection time. On the contrary, only a subpopulation of AT2 cells (about 30-40%) was sensitive for viral infection although they could be quickly infected (within 24 hpi). The identity of the SARS-CoV-2 sensitive AT2 cell subpopulation and why other AT2 cells could not be infected need further investigation.

We noted that viral infected cells expressed ACE2 but not all ACE2 expressing cells were infected. TMPRSS2 is another known factor that determines SARS-CoV-2 cell entry 5 , and therefore we checked the expression pattern of TMPRSS2 in human lung organoids. Immunostaining analyses showed that TMPRSS2 was ubiquitously expressed in both hAWOs and hALOs, which was contrary to the restricted expression pattern of ACE2 (Fig. S2 ).

Therefore, compared to TMPRSS2, ACE2 was the major factor that determined the cell tropism of SARS-CoV-2 in human lung organoids.

Next, we checked whether SARS-CoV-2 infection was associated with proliferation status by co-immunostaining with viral N protein and Ki67 (cycling marker). We found that infected cells (NP + ) contained both cycling (Ki67 + ) and noncycling (Ki67 -) cells in hAWOs and most infected cells were cycling cells in hALOs (Fig. S3a) . We then checked whether SARS-CoV-2 infection induced apoptosis by co-immunostaining with viral N protein and cleaved Caspase3 (C-Caspas3, apoptotic cell marker). No obvious cell death was observed at 24 hpi or 48 hpi, but at 72 hpi, apoptosis became prominent in both organoids, particularly more in hALOs (Fig. S3b-d) .

To conform the viral replication, the ultrastructures of infected hAWOs and hALOs were analyzed by transmission electron microscopy at 72 hpi or 96 hpi.

Part of hAWOs and hALOs in one mesh of the grids were shown in Fig. 4a and 4e, and viral particles were found in cells of both organoids ( Fig. 4b -d, f and g).

In both organoids, viral particles were observed in the apical, lateral and basolateral side of the cells (Fig. 4h-j) , indicating potential dissemination route how SARS-CoV-2 passes across the lung epithelial barrier. Double membrane vesicles (DMVs) and convoluted membranes (CMs) with spherules are typical coronavirus replication organelles 42,43 , which were observed in the lung organoids ( Fig. 4k ). Virus particles in cells were seen in membrane bound vesicles, either as single particles or as groups in enlarged vesicles (Fig. 4l ).

Enveloped viruses were observed in the lumen of Golgi apparatus and secretory vesicles (Fig. 4m, n) , which was consistent with previous report that coronaviruses assembled and matured at the endoplasmic reticulum-Golgi (Fig. 4b, g and n) . Besides, virus particles were found in late endosomes with engulfed cell debris (Fig. 4p, q) . And more dying cells and engulfed cell debris were observed in hALOs than in hAWOs (Fig. 4r) . The TEM data ( Fig. 4p-r) , as well as the C-Caspase3 immunostaining data ( Fig. S3b-d) , indicated that the pathological changes of alveoli and bronchioles after SARS-CoV-2 infection were different.

To determine the early cell response to SARS-CoV-2 infection, we performed RNA-sequencing analysis using hAWOs and hALOs 48h after SARS-CoV-2 infection. Abundant SARS-CoV-2 viral RNA was detected solely in the infected organoids (Fig. 5a) . Principle component analysis (PCA) showed that the samples formed four separate clusters according to organoid type and virus infection (Fig. 5b ). In total, 1679 differential expressed genes were ). ACE2 is the receptor for SARS-CoV and SARS-CoV-2, and SARS-CoV spike (S) protein can induce shedding of ACE2 by ADAM17, which is believed to be a crucial mechanism for SARS-CoV-induced lung injury [49] [50] [51] [52] . Surprisingly, we found that the mRNA expression level of ACE2 was downregulated at 48h after SARS-CoV-2 infection (Fig. 5f ). Since most infected cells were viable at 48 hpi ( Fig. S3b-d) , the downregulation of ACE2 mRNA was not a secondary effect of cell death but a direct effect of virus infection. Therefore, we believe that SARS-CoV-2 infection might decrease the expression of ACE2 at both protein and mRNA levels. However, the mechanisms of downregulation remain open for further investigation. In addition, we found that the expression of TMPRSS2 was also slightly downregulated after SARS-CoV-2 infection at a much less extent than ACE2 (Fig. 5f ).

Finally, we tested the inhibitory effect of remdesivir, camostat, and bestatin on the infection of human lung organoids by SARS-CoV-2. Remdesivir is a nucleotide analogue prodrug to inhibit viral replication 53 , which has been reported to repress SARS-CoV-2 infection in basic research and clinic trials 12, 54, 55 . Camostat is an inhibitor of the serine protease TMPRSS2 that cleaves SARS-CoV-2 S protein and facilitates viral entry 5 . Bestatin is an inhibitor of CD13 (Aminopeptidase N/APN) 56 , a receptor utilized by many αcoronaviruses (SARS-CoV-2 belongs to β-coronaviruses) 44 . As shown in Fig.   6a , remdesivir reduced the production of infectious virus in hAWOs and hALOs, and camostat showed a slightly inhibitory effect in hAWOs not in hALOs, while bestatin had no effects in either hAWOs or hALOs. Quantitative RT-PCR analyses of supernatant viral RNA also demonstrated that remdesivir inhibited viral load (Fig. 6b ). We noted that remdesivir reduced viral load to 1/10 but infectious virus titer to less than 1/1000. Similar phenomena, with potent inhibitory effect on virus titer and much less effect on viral load, have been reported in remdesivir treated rhesus macaques with SARS-CoV-2 infection 16 .

An explanation for the phenomena might be that virus particles with RNA containing the remdesivir-metabolized adenine analogue are defective for infection, in addition to the known mechanism that remdesivir induces delayed chain termination 53 .

In summary, we demonstrated that hESCs-derived airway and alveolar organoids could be infected by SARS-CoV-2 and be used for drug test, serving as a pathophysiological model to complement cell lines and animals.

Medicine and Health Guangdong Laboratory (2020GZR110106006 to X.C. and J.C.), the emergency grants for prevention and control of SARS-CoV-2 of Guangdong province (2020B111108001 to X.C.)

and National Postdoctoral Program for Innovative Talent (BX20190089 to X.Y.)

X.C., Z.R., and J.C. initiated, designed and supervised this study; R.P. 

All experiments in the present study were performed on H9 human embryonic stem cells (hESCs). hESCs were maintained in feeder-free culture conditions in 6-well tissue culture dishes on Matrigel (BD Biosciences, 354277) in mTeSR1 medium (Stem Cell Technologies, 05850) at 37°C with 5% CO2.

Cells were passaged with TrypLE (Gibco) at 1:6 to 1:8 split ratios every 4 days.

hESCs derived hAWOs and hALOs were generated as previously Supplementary Table S1 .

SARS-CoV-2 (WIV04) 4 cells.

Total RNA in the cells was extracted using Trizol (Invitrogen, 15596026) according to the manufacturer's protocol, and 1ug RNA was used to reverse Reads were aligned to the human reference genome hg38 with bowtie2 5 , and RSEM 6 was used to quantify the reads mapped to each gene. Gene expression was normalized by EDASEQ 7 . Differentially expressed genes were obtained using DESeq2 (version 1.10.1) 8 , a cutoff of Q-value < 0.05 and log2

(fold-change) > 1 was used for identify differentially expressed genes. All differentially expressed mRNAs were selected for GO analyses clusterProfiler 9 .

Other analysis was performed using glbase 10 . The RNA-seq supporting this study is available at GEO under GSE155717. Data are accessible with a reviewer token: "mbcxaucmpbwttup".

For immunofluorescence staining, samples were transferred into 1.5ml tubes and fixed with 4% paraformaldehyde overnight at 4°C or 2h at RT. Supplementary Table S2 .

Organoids were collected and fixed in 2.5% glutaraldehyde for 24h, washed with 0.1M Phosphate buffer (19ml 0.2M NaH2PO4, 81ml 0.2 M Na2HPO4) for 3 times, and further fixed with 1% Osimium tetraoxide for 2h at room temperature.

The fixed organoids were then washed with phosphate buffer and dehydrated with 30%, 50%, 70%, 80%, 85%, 90%, 95%, and 100% alcohol sequentially. 

The epidemic of 2019-novel-coronavirus (2019-nCoV) pneumonia and insights for emerging infectious diseases in the future

Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro

The pathogenicity of SARS-CoV-2 in hACE2 transgenic mice

Pathogenesis of SARS-CoV-2 in Transgenic Mice Expressing Human Angiotensin-Converting Enzyme 2

ChAdOx1 nCoV-19 vaccination prevents SARS-CoV-2 pneumonia in rhesus macaques. bioRxiv

Clinical benefit of remdesivir in rhesus macaques infected with SARS-CoV-2

SARS-CoV-2 infection protects against rechallenge in rhesus macaques

Comparative pathogenesis of COVID-19, MERS, and SARS in a nonhuman primate model

Susceptibility of ferrets, cats, dogs, and other domesticated animals to SARScoronavirus 2

DNA vaccine protection against SARS-CoV-2 in rhesus macaques

Pathogenesis and transmission of SARS-CoV-2 in golden hamsters

Mice are not men

Long-term expansion of alveolar stem cells derived from human iPS cells in organoids

Efficient Derivation of Functional Human Airway Epithelium from Pluripotent Stem Cells via Temporal Regulation of Wnt Signaling

Long-Term Engraftment Promotes Differentiation of Alveolar Epithelial Cells from Human Embryonic Stem Cell Derived Lung Organoids

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Fast gapped-read alignment with Bowtie 2

RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome

GC-content normalization for RNA-Seq data

Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

clusterProfiler: an R package for comparing biological themes among gene clusters

glbase: a framework for combining, analyzing and displaying heterogeneous genomic and high-throughput sequencing data

We thank Prof. Mengfeng Li from Southern Medical University for helpful discussion. We are particularly grateful to Tao Du, Lun Wang and the running