Modeling COVID-19 with Human Pluripotent Stem Cell-Derived Cells Reveals Synergistic Effects of Anti-inflammatory Macrophages with ACE2 Inhibition Against SARS-CoV-2

Dysfunctional immune responses contribute critically to the progression of Coronavirus Disease-2019 (COVID-19) from mild to severe stages including fatality, with pro-inflammatory macrophages as one of the main mediators of lung hyper-inflammation. Therefore, there is an urgent need to better understand the interactions among SARS-CoV-2 permissive cells, macrophage, and the SARS-CoV-2 virus, thereby offering important insights into new therapeutic strategies. Here, we used directed differentiation of human pluripotent stem cells (hPSCs) to establish a lung and macrophage co-culture system and model the host-pathogen interaction and immune response caused by SARS-CoV-2 infection. Among the hPSC-derived lung cells, alveolar type II and ciliated cells are the major cell populations expressing the viral receptor ACE2 and co-effector TMPRSS2, and both were highly permissive to viral infection. We found that alternatively polarized macrophages (M2) and classically polarized macrophages (M1) had similar inhibitory effects on SARS-CoV-2 infection. However, only M1 macrophages significantly up-regulated inflammatory factors including IL-6 and IL-18, inhibiting growth and enhancing apoptosis of lung cells. Inhibiting viral entry into target cells using an ACE2 blocking antibody enhanced the activity of M2 macrophages, resulting in nearly complete clearance of virus and protection of lung cells. These results suggest a potential therapeutic strategy, in that by blocking viral entrance to target cells while boosting anti-inflammatory action of macrophages at an early stage of infection, M2 macrophages can eliminate SARS-CoV-2, while sparing lung cells and suppressing the dysfunctional hyper-inflammatory response mediated by M1 macrophages.


Introduction
The infection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has already caused more than 5.4 million Coronavirus Disease-2019 (COVID-19) cases internationally (https://google.org/crisisresponse/covid19-map). Most COVID-19 patients show mild to moderate symptoms of fever, dry cough, fatigue and diarrhea, however, approximately 15% of con rmed cases progress to severe pneumonia, acute respiratory distress syndrome (ARDS) or multi-organ failure (Guan et al., 2020). The progression from mild to severe disease or death is principally attributed to dysfunctional immune responses (Mehta et al., 2020;Wang et al., 2020) together with viral damage of target cells. Given the lack of an effective vaccine or medication, a thorough understanding of immunological features caused by SARS-CoV-2 is critically important for studying viral pathobiology and therapeutic development.
Alveolar macrophages (AMs) are key sentinel cells for host defense in the respiratory system, producing cytokines and chemokines that are crucial components of innate immunity and mediators of immunopathology (Allard et al., 2018). The polarization of macrophages confers a heterogeneous function and plasticity depending on the duration of stimulation and microenvironment, which are discrete phenotypes associated with different in ammatory responses, typically termed the M1φ /pro-in ammatory and M2φ /anti-in ammatory macrophages (Gomez Perdiguero et al., 2015;Wynn et al., 2013). The distinction is known to be oversimpli ed, with macrophage dynamic activities spread along the M1-M2 phenotypic spectrum (Bian, 2020;Shapouri-Moghaddam et al., 2018). However, in general, M1φ destroys pathogens by producing a large number of pro-in ammatory cytokines such as IL-1β, TNFα, IL-6 and IL18. In contrast, M2φ exhibits higher activity in phagocytosis against pathogens and for anti-in ammation (Mills, 2015;Murray, 2017). Recent studies (Liao et al., 2020;Xu et al., 2020) on immunity of COVID-19 patients indicate that the cells damaged by SARS-CoV-2 infection induced innate in ammation in the lungs that is largely mediated by pro-in ammatory macrophages and granulocytes. In addition to local damage, the pro-in ammatory macrophages release cytokines/chemoattractants and prime adaptive immune cell responses, which in some cases lead to dysfunctional immune responses and cytokine storm, followed by respiratory and even multi-organ failure . These studies imply a crucial role for macrophages in the progress of SARS-CoV-2 infection; a deeper understanding of the interactions among targeted cells, macrophages and SARS-CoV-2 could offer new ideas to help combat this deadly contagious disease.
The current most widely used model for SARS-CoV-2 research is the African green monkey derived Vero cells, which are very limited for modeling human disease. Although primary macrophages are more functionally or phenotypically representative of native macrophages in the tissue from which they are derived, they are di cult to obtain, proliferate slowly, and are often poorly characterized (Jobe et al., 2017). In this study, we generated lung cells and macrophages paired from the same cell origin, human pluripotent stem cell (hPSC) lines. This strategy overcomes a common concern about histocompatibility when studying human immune cells with other cell types, and provides theoretically unlimited cell resources for reliably modeling and studying immunology of macrophages and human lungs during SARS-CoV-2 infection. Our results using this platform demonstrate a potential therapeutic strategy through a combination of boosting anti-in ammatory macrophages and intervention of viral entry, to control SARS-CoV-2 infection at the immune defense-based protective phase while circumventing the in ammation-driven damaging phase.

Results
Macrophage involved at the severe stage of COVID-19 To better understand how macrophages impact COVID-19 progression, we compared immune cells and in ammatory factors in lung tissues obtained from autopsies of COVID-19 patients or healthy donors. First, histological changes in lung tissues from COVID-19 patients were examined. Compared to healthy lung tissues, this revealed extensive necrotizing bronchiolitis with necrotic bronchial epithelial cells and severe alveolitis with atrophy and desquamation, displayed in the lumen of the patient's lung ( Figure 1A). Of note, pulmonary hemorrhagic infarct with abundant in ammatory in ltration (arrow heads) were extensively present through the whole alveoli and bronchial regions ( Figure 1A). Recently, it was reported that proin ammatory FCN + monocyte-derived macrophages were mainly present and FABP4 + alveolar macrophages were greatly reduced in the bronchoalveolar lavage uid from patients with severe COVID-19, whereas mild and moderate cases were characterized by the presence of highly clonally expanded CD8 + T cells (Liao et al., 2020). Therefore, we examined if macrophages were dominantly present in the diseased patient's lung. Immunostaining against pan macrophage marker CD68 showed abundant macrophages were extensively distributed through the whole lung tissue with aggregated phenotypes (Figure 1B), in agreement with the above-cited report. However, macrophages are multifaceted and distinct functions of macrophages highly depends on polarization, characterized generally as M1/pro-in ammatory or M2/anti-in ammatory macrophages. We thus further examined M1 macrophage marker CD80, and M2 macrophage marker CD163 ( Figures 1C-F). The results revealed that cells positive for either CD80 or CD163 were both aberrantly represented in the patient's lung tissue (Figures 1C-F). Indeed, CD68 + , CD80 + and CD163 + macrophage populations were signi cantly expanded in the patient's lung tissue, suggesting expansion of both M1 and M2 macrophage populations in severe disease. We also examined several cytokines that are mainly produced by macrophages and found key pro-in ammatory cytokine IL-6 was intensively expressed in the lumen of the patient's lung tissue ( Figure 1G). Taken together, the data supports a need to further examine the roles of M1 and M2 macrophages in COVID-19 progression.

Co-culture of lung cells and macrophages derived from hPSCs
To further investigate the interaction among macrophages, lung cells and SARS-CoV-2, we established a co-culture model using cells derived from the same hPSC line (RUES2 or H1), which provide a genetically de ned background for immune study.  (Figures S3A-D). ACE2, the putative SARS-CoV-2 receptor, and TMPRSS2, the co-effector for viral entry (Hoffmann, 2020), were detected in AT2, AT1 and ciliated cells, in clusters 0, 2, 3 ( Figures S3E-F). The immunostaining results further validated that ACE2 is mainly co-expressed with SP-B or pro-SP-C in AT2 cells, and FOXJ1 in ciliated cells ( Figure S2B), consistent with results previously reported in primary human lung tissues (Ziegler, 2020).
Next, the hPSC-induced lung cells (iLung) and macrophages (iMφ) were plated and cultured together in a 1:1 ratio (Figure 2A), similar to the ratio of lung cells and macrophages in distal bronchial or alveolar regions in human lung (Kyle J. Travaglini, 2020). The iLung was derived from the hPSC lines carrying a Doxycycline-inducible GFP reporter gene, which allowed the distinction of iLung and iMφ in live cultures ( Figure 2B). A signi cantly lower number of GFP + iLung were observed after four-day co-culture with iMφ of M1 phenotype (iM1φ), than seen in the co-culture with iMφ of M2 phenotype (iM2φ) or control 293T cells ( Figure 2C). The scRNA pro les further revealed decreased expression of proliferation-associated genes MKI67 and TOP2A and increased expression of apoptosisrelated genes TP53, CASP3, BAX, MCL1, in the iLung co-cultured with iM1φ, but not in co-cultures with iM2φ ( Figure S6D). These results were in alignment with the phenotype of pro-in ammatory activities of iM1φ, as scRNA-seq data detected a set of proin ammatory factors, IL1B, IL18, STAT1, FCN1, CXCL9, CXCL10, CXCL11, CXCL16, CCL2 highly expressed in iM1φ ( Figure 2F-G, S5B-C). In contrast, iM2φ mainly expressed anti-in ammatory factors or immunoregulatory genes such as TGM2, APOE, A2M, CCL13, CCL26 and TREM2 ( Figure 2F-G, S5B-C). Gene Ontology (GO) enrichment analysis comparing iM1φ and iM2φ revealed overactivation of differential signaling pathways such as pro-in ammatory IFNγ, type I IFN, and neutrophil activation in iM1φ; antiin ammatory and tissue damage-repair process of RNA catabolic process, protein co-localization to endoplasmic reticulum in iM2φ ( Figures S6B, C). Similar phenotypes were observed in the iLung co-cultured with THP-1, an established monocyte line, upon activation of M1 or M2 phenotype ( Figure 2C). The results indicate that activation of M1-macrophage was su cient to create a toxic environment for the iLung even in the absence of viral infection.

Immune response of macrophages following SARS-CoV-2 infection
To model the immune response of macrophages to SARS-CoV-2 infection on lung cells, virus was introduced to the co-culture system ( Figure 3A). As a rst step to measure effects of macrophages on viral entry into lung cells, we used a SARS-CoV-2 pseudoentry virus, in which the backbone of a VSV-G pseudo-typed ΔG-luciferase virus carries the SARS-CoV-2 spike protein incorporated in the surface of the viral particle (Nie et al., 2020;Whitt, 2010). High luciferase activity was readily detected in iLung 24 hours after viral infection at MOI = 0.01, but not in iMφ or 293T in the co-culture (293T cells were used as a co-culture control, based on our preliminary data and previous report that the permissiveness of 293T to SARS virus is low (Wenhui Li, 2003)) ( Figure 3B), and immunostaining con rmed that the viral luciferase protein was co-localized with ACE2 + cells in the iLung cultures ( Figure S7B). Since the luciferase gene was expressed after the virus entered host cells, the luciferase activity correlated to the amount of viral entry host cells. Luciferase activity was markedly decreased in the co-cultures of iLung with all three lines of macrophages, iMφ, THP-1 and U937; no signi cant difference was found between hPSC-derived iM1φ or iM2φ, indicating they have the similar inhibitory effects on viral infection ( Figure 3B, Figure S7A). The results were further validated by immunostaining study that substantial decrease of luciferase protein was detected in iLung cells co-cultured with iMφ, compared to those co-cultured with 293T ( Figure S7A). The potential of iMφ to inhibit viral replication and spreading was next studied by infection with a patient-derived SARS-CoV-2 virus in the co-cultures. After 24 hours incubation with the SARS-CoV-2 virus (USA-WA1/2020, MOI = 0.01), a signi cant decrease of viral protein was observed in the co-culture of iLung and iMφ, compared to the co-culture of iLung and 293T. Strikingly, most SARS-CoV-2 virus SARS-N protein was detected in the M2-iMφ when co-cultured with iLung, while in contrast, substantial levels of SARS-N protein was detected in iLung cells in the co-cultures using M1-iMφ or 293T ( Figure 3D). The ndings suggest that phagocytosis activity of M2-iMφ functioned as protection for iLung from viral infection.
Several approaches were taken to thoroughly examine the immune response following iMφ on SARS-CoV-2 infection. First, a cohort of cytokines and in ammatory factors that are known to be important for innate or adaptive immune responses were pro led, in the culture medium 24 hours after infection with the SARS-CoV-2 pseudo virus. Increased levels of IFNγ, IL-6, and IL-18 were found in the co-cultures of iLung with M1-iMφ, while these were decreased in the co-cultures of iLung with M2-iMφ ( Figure 3C). To further characterize at the transcriptomic level the response of iLung and iMφ following viral infection, scRNA-seq was performed on the cocultures with SARS-CoV-2 pseudo virus infection and the analysis revealed that a set of anti-in ammatory factors and anti-viral activity related genes, such as CCL26, CCL13, ISG15, IFITM2 and IFITM3, were clearly upregulated when cultures contained M2-iMφ ( Figure 4A and C, Figures S8A). In contrast, pro-in ammatory factors, such as IL-6, S100A8/A9, LYZ and TLR4 were highly expressed when the cultures contained M1-iMφ. ( Figure 4A, C, Figures S8A). Gene enrichment analysis comparing iM1φ and iM2φ revealed over-activation of differential signaling pathways such as neutrophil degranulation and antigen processing and presentation, regulation of T cell mediated cytotoxicity in iM1φ; granulocyte chemotaxis, response to interferon−gamma as well as phagocytosis in iM2φ ( Figures 4 E and F). Moreover, IL10 signaling related genes such as IL10RA, IL10RB, STAT3, SOCS3, TIMP1 and IRS2 were enriched in iM2φ, suggestive of anti-in ammatory macrophages ( Figure S8D). The above results demonstrate a differential immune response of iM2φ versus iM1φ upon viral entry into host cells, as iM2φ increased phagocytosis activity and released antiin ammatory factors, while iM1φ increased antigen-presenting activity and released pro-in ammatory factors.
Correlating with the above phenotypes, up-regulation of cell growth arrest or death-related genes, such as GAS6, BTG2, PDCD6, CCAR1, TP53I11, TP53INP1, and activation of programmed death signaling pathways as well as higher mitochondrial genes,MT−CYB,MT−CO1, MT-CO2, MT-ND1 ( Figure S8B and C), were detected in the co-cultures with iLung with iM1φ, but not with iM2φ ( Figure S8B). Previous studies by us (Yuling Han, 2020) and others (Conti et al., 2020) suggested that lung cells display selfimmune defense after SARS-CoV-2 infection, releasing proin ammatory factors, such as CXCL2, CCL2, CXCL3 and IL1A, as well as BCRC3, AADAC, and ATPB4. The GO and KEGG analysis in our current co-culture based data suggest that upregulation in pathway networks including leukocyte chemotaxis NF-κB signaling, IL-17 signaling, viral protein interaction with cytokine-cytokine receptor, and response to type I interferon, combined with the pro-in ammatory reaction of M1 macrophages, could lead to further pulmonary in ammation and damage ( Figure S8E). Moreover, the scRNA-seq pro ling data further validated the immunostaining results showing that few if any iLung cells in the co-culture with M2-iMφ displayed detectable viral gene expression, in contrast to a signi cantly higher number of iLung cells in the co-culture with M1-iMφ ( Figure 4B). Most infected AT2 cells and ciliated cells were also found in the co-culture with M1-iMφ, indicating a stronger protective effect on iLung cells by M2-iMφ ( Figure 4D). Altogether, these ndings suggest that activation of pro-in ammatory macrophages can aggravate lung cell damage, beyond the destruction by viral infection; in contrast, activation of anti-in ammatory macrophages provides a protective effect for lung cells from viral infection.

Blockage of ACE2 enhances elimination of SARS-CoV-2 by macrophages
Several studies (Tay et al., 2020) on mild or recovered COVID-19 cases indicated that in a healthy immune response, neutralizing antibodies produced in these individuals can block viral infection, followed by alveolar macrophages recognizing the neutralized viruses and clearing them by phagocytosis. We sought to model this process using an ACE2 blocking antibody to inhibit virus entry to target cells, thus decreasing the viral loads ( Figure 5A), to test if this enhances phagocytosis activity of macrophages. As expected, incubation with ACE2 blocking antibody two hours prior to infection of SARS-CoV-2 pseudo virus, reduced markedly the luciferase activities in co-cultures of iLung with either M1 or M2-iMφ, although the decrease of luciferase signal was most pronounced in the cocultures with M2-iMφ ( Figure 5B, Figure S9A). Immunostaining results validated that luciferase protein dramatically decreased in the iLung cells co-cultured with iMφ, compared to those co-cultured with 293T ( Figure S9A).
Similarly, immunostaining results from the experiments performed using SARS-CoV-2 virus further revealed that that most SARS-CoV-2-N protein was found in the M2-iMφ, but not in the iLung cells, while the N protein was clearly found in iLung cells in the coculture with M1-iMφ or 293T ( Figure 5C). These results demonstrated that an early intervention of viral infection by blocking ACE2 in target cells can increase the clearance of virus by macrophages, especially synergizing with the phagocytosis activity of M2macophages to further provide protection for target cells and reduce the damage by in ammatory factors produced by M1macrophages.

Discussion
The study of human host-immune systems with pathogens has depended historically on the use of animal models, largely due to limited cell resources derived from human tissues. Immune research on COVID-19 is limited by the types of models available for study. Recently, a transgenic mouse strain(McCray et al., 2007) has been made with human ACE2 expression regulated by human cytokeratin-18 promoter, but the ACE2 expression in human is more complex than that in the mice. Another model is ferret (Kim et al., 2020), which can be infected with SARS-CoV-2, but does not develop hyper-in ammation in the lung. Recent advances in stem cell biology, especially the technology to differentiate human pluripotent stem cells (hPSCs) into functional immune cell types, provide a rigorous human system for studying immunology and therapeutics. In this report, we describe a new cell co-culture system in which the immune cells, speci cally monocytes/macrophages, and lung lineage cells are produced by directed differentiation of hPSCs.
Several key features make the human cell model an ideal system for studying immunology of SARS-CoV-2. The model contains the host cells and immune cells from the same hPSC lines, avoiding concern of histocompatibility, while it can provide abundant numbers of cells with a genetically de ned background for robust mechanistic or therapeutic studies.
The innate immune response mainly mediated by macrophages or granulocytes, responding to tissue damage caused by SARS-CoV-2 infection, likely contributes to acute respiratory distress syndrome (ARDS) that is characterized by the rapid onset of widespread in ammation in lung and subsequent respiratory failure . Our study in COVID-19 patient samples validated a correlation of macrophages and the disease, showing a heavy in ltration of pro-in ammatory macrophages in tissue samples from distal lung regions with high levels of in ammatory cytokine IL-6 in the severe cases. The macrophage and lung cell co-culture model combined with single cell transcriptomics was then applied to interrogate the differential immune responses of proor anti-in ammatory macrophages following SARS-CoV-2 infection. We discovered that pro-and anti-in ammatory macrophages both have similar capacity to eliminate the virus in the context of a moderate viral load. However, the immune reaction of proin ammatory macrophages following SARS-CoV-2 pseudo-virus infection led to more damage on lung cells and secretion of a set of in ammatory factors including IL6, IL18 and CXCL10 that are known to be mediators in dysfunctional immune responses and cytokine release syndrome (CRS). In contrast, anti-in ammatory macrophages protected lung cells from viral infection, and diminished pulmonary in ammation by phagocytosis and production of anti-in ammatory factors related in IL10 signaling. Finally, inhibiting viral entry in target cells using an ACE2 blocking antibody, diminished viral infection and enhanced the elimination of viruses. In particular, the intervention on viral entry can synergize with the phagocytosis and antiviral activity of macrophages, resulting in a more pronounced clearance of virus and protection of target cells.  The para n-embedded lung tissues were acquired from the department of pathology in the 3 rd people's hospital of Shenzhen, China. They recently reported the pathological changes of lungs from a 66-year-old male died in critical COVID-19 infection(Weiren Luo). The patient developed respiratory failure and septic shock during the treatment and was done with transplant. Informed consent was obtained from the patient and family. The diagnosis of COVID-19 pneumonia was based on the "Coronavirus Pneumonia Prevention and Control Plan" (7th edition) newly issued by the National Health Commission, China (Commission, 2020). Nasopharyngeal swabs were collected and COVID-19 was detected by real-time polymerase chain reaction. Infection was de ned as at least two positive test results. Surgical informed consent was obtained and the study was approved by IRB in the third People's hospital of Shenzhen. . For macrophage differentiation, at day -2, hESCs were digested into single-cell suspension by 1 mg/ml Accutase (Stemcell Technologies) and plated onto Matrigel-coated culture dishes at a density of 2× 10 4 cells/cm 2 in mTeSR1 medium with 5uM Y27632 (MedchemExpress). After 24 h, Y27632 was withdrawn from the medium and cells were cultured for another 24 h. At day 0, cells were rstly induced by macrophage differentiation basal medium (SFD-M) which is RPMI 1640 medium supplemented with 2% B27 (Thermo Fisher Scienti c), 1% L-GlutaMAX-I and 50 μg/ml ascorbic acid (Sigma Aldrich) and 10 ng/ml BMP4 (R&D Systems) for 24 h. Afterward, the medium was changed to SFD-M medium containing 10 ng/ml BMP4 and 2 μM GSK3 inhibitor CHIR99021 (Cayman Chemical) for another 48 h. At day 3, cells were replated onto Matrigel-coated dishes at a density of 4 × 10 4 cells/ cm 2 in SFD-M medium with 50 ng/ml VEGF (R&D Systems) and 10 ng ng/ml FGF2 (R&D Systems) for 48 h. At day 5, the medium was replaced with basal medium with 50 ng/ml VEGF, 10 ng ng/ml FGF2 and 10uM TGFβ inhibitor SB431542 (R&D Systems) for another 72 h. At day 8-10, oating cells were collected and medium was changed and supplemented with 50ng/ml M-CSF and 10ng/ml IL3 (R&D Systems) for another 3-5 days. From day 11-13 onward, the medium was changed to SFD-M medium with 50 ng/ml M-CSF for 3 days. All differentiation steps were cultured under normoxic conditions at 37 ℃, 5% CO 2 . The protocol details are summarized in Figure S4A.
All embryonic stem cell studies were approved by the Institutional Review Board (IRB) at the University of Chicago, or by the Tri-Institutional ESCRO committee (Weill Cornell Medicine, Memorial Sloan Kettering Cancer Center, and Rockefeller University).
hPSC monocyte polarization hPSC-derived CD14 + cells were plated on tissue culture plates at a density of 2x10 4 cells/cm 2 in SFD-M medium supplemented with 50 ng/mL M-CSF. After 2 days of culture, monocytes differentiated into M0 macrophages and polarized to M1 or M2 macrophages. For macrophages polarization, 100ng/mL LPS (Sigma-Aldrich) and 10ng/mL IFNγ (R&D Systems) were added for M1 induction, or 20 ng/m IL-4 (R&D Systems) was added for M2 induction in SFD-M medium supplemented with 50 ng/mL M-CSF, respectively. These cells were cultured for another three days before examination for expression of the M1 or M2 makers.

Giemsa Staining
Differentiating day 11-13 monocytes/macrophages were xed on slides using Cytospin, followed by staining using Wright-Giemsa Stain (Sigma-Aldrich) according to the manufacturer's instructions.

Immunohistochemical staining
Histological study of lung tissues was performed on para n-embedded sections as previously described (Li et al., 2014). For immunohistochemical staining, para n-embedded sections were depara nized and incubated with primary antibodies at 4°C overnight and secondary antibodies at room temperature for 1h. Primary antibodies and secondary antibodies are described in the supplementary Table. Nuclei were counterstained by Hoechst 33342 (Sigma). positive cells in lungs were randomly counted from For FACS analysis, cells were resuspended in a FACS buffer (PBS with 0.1 % BSA and 2.5 mM EDTA). The cell suspension was then stained with PE-conjugated CD43 (Biolegend, clone MEM-59), APC-conjugated CD34 (BD, clone 581) to detect hematopoietic stem/progenitor cells (HSPC). PE-conjugated CD68 (Biolegend, clone Y1/82A), APC-conjugated CD11b (Biolegend, clone ICRF44), FITC-conjugated CD14 (Biolegend, clone HCD14) were used to detect monocyte/macrophages. Basically, cells were incubated with antibodies for 30 minutes at 4°C, followed with washed and suspended in 0.1% BSA/PBS buffer. PE and APC lters were then used to detect cells double positive for CD43 and CD34 or CD68 and CD11b by signal intensity gating, FITC and APC were used to detect cells double positive for CD14 and CD11b. Negative controls stained with control IgG instead of primary antibodies were always performed with sample measurements. Flowcytometry machine of BD FACSAria II and software of Flowjo were mainly used to collect and analyze the owcytometry data.

SARS-CoV-2-Pseudo-Entry Viruses
Recombinant Indiana VSV (rVSV) expressing SARS-CoV-2 spikes was generated as previously described (Nie et al., 2020;Whitt, 2010;Zhao et al., 2017). HEK293T cells were grown to 80% con uency before transfection with pCMV3-SARS-CoV2-spike (kindly provided by Dr. Peihui Wang, Shandong University, China) using FuGENE 6 (Promega). Cells were cultured overnight at 37°C with 5% CO 2 . The next day, the media was removed and VSV-G pseudotyped ΔG-luciferase (G*ΔG-luciferase, Kerafast) was used to infect the cells in DMEM at an MOI of 3 for 1 hr before washing the cells with 1X DPBS three times. DMEM supplemented with 2% FBS and 100 I.U. /mL penicillin and 100 μg/mL streptomycin was added to the infected cells and they were cultured overnight as described above. The next day, the supernatant was harvested and clari ed by centrifugation at 300xg for 10 min before aliquoting and storing at −80°C. To assay pseudo-typed virus infection, cells were seeded in 96 well plates. Pseudo-typed virus was added for MOI=0.01. At 2 hpi, the infection medium was replaced with fresh medium. At 24 hpi, cells were harvested for luciferase assay or immunohistochemistry analysis. For liver and lung organoids, organoids were seeded in 24-well plates, pseudo-typed virus was added for MOI=0.01 and centrifuged the plate at 1200g, 1 hour. At 24 hpi, organoids were xed for immunohistochemistry or harvested for luciferase assay following the Luciferase Assay System protocol (E1501, Promega)

SARS-CoV
Single cell sequencing of hPSC-derived lung cells Single-cell capture, reverse transcription, cell lysis, and library preparation was performed using the Single Cell 3′ version 3 kit and chip according to the manufacturer's protocol (10x Genomics, USA). Single-cell suspensions were generated by dissociating the cultured RUES2 cells with 0.05% Trypsin/0.02% EDTA for 10-15 min, followed with passing through 40µM strainer. The single cell suspension was achieved through sorting the dissociated cells in ow cytometry singlets. Cell count was adjusted to 1000-2000 cells per ul to target an estimated capture of 8000 cells. Six input wells were used. Sequencing was performed on NovaSeq6000 with setting 28 for read 1 and 91 for read 2. The sequencing data were primarily analyzed by CellRanger pipeline v3.0.2 (10x Genomics, USA). In particular, raw fastq data were generated by CellRanger mkfastq; A custom reference genome was built by integrating the virus and luciferase sequences into the 10x pre-built human reference (GRCh38 v3.0.0) using CellRanger mkref. Alignment of the raw reads to the custom reference genome, removing duplicated transcripts using the unique molecular identi ers (UMIs) and assignment to single cells was performed using CellRanger count.
Brie y, we used cells Seurat 3.1.4 R package for data analysis and visualization (Butler et al., 2018). The Seurat object is required at least 200 and at most 6000 unique molecular identi ers (UMIs), genes detected (UMI count > 0) in less than two cells were removed. In addition, cells were excluded if more than 10% of sequences mapped to mitochondrial genes. In total, 5,080 cells from the sample passed these lters for quality.
Following the package suggestions, we used a linear model to mitigate the variation stemming from the number of detected unique molecules per cell. The differentially expressed genes were found by ''vst'' method and the top 3,000 differentially expressed genes were selected for PCA analysis. We used an elbow plot to determine the number of PCs. 20 PCs were used for each group of cells. Clustering resolution was set at 0.2. For co-culture analysis, Macrophages and lung cells were re-clustered and re-analyzed, respectively. Macrophages were integrated using the rst 20 dimensions of PCs and clustering resolution was set at 0.  DATA AVAILABILITY scRNA-seq data is available from the GEO repository database with accession number GSE150708 (hPSC-derived lung cells, Coculture of macrophage and lung cells derived from hPSC, Co-culture of macrophage and lung cells derived from hPSC in SARS-CoV-2 infection).
Competing Financial Interests The authors have no nancial con icts of interest. Figure 1 Macrophages were highly involved at the severe stage of COVID-19 (A) H+E (Hematoxylin and Eosin) staining on the bronchial or alveolar region in healthy or severe COVID-19 case. Pulmonary hemorrhagic infarct (denoted by arrow heads) (B) Immunohistochemistry (IHC) using antibody against CD68 revealed macrophage with aggregated phenotype and enlarged nuclei in COVID-19 lung, compared to the ones in healthy lung. (C) Immuno uorescence (IF) staining on healthy or COVID-19 distal lung tissues using antibodies against CD68 (pan-macrophage marker), and CD80 (M1 macrophage marker) (D) Quanti cation on CD68+ or CD80+ macrophages in healthy or COVID-19 distal lung tissues. (E) IF staining on healthy or COVID distal lung tissues using antibodies against CD68 and CD163 (M2 macrophage marker) (F) Quanti cation on CD68+ or CD163+ macrophages in healthy or COVID-19 distal lung tissues. (G) IF staining on healthy or COVID-19 distal lung tissues using antibodies against CD68 and IL-6. Scale bar = 100 µm in all images in Figure 1. Data was presented as mean ± STDEV. P values were calculated by unpaired two-tailed Student's t test. **P < 0.01, ***P < 0.001, and ****P < 0.0001.  The effects of macrophages in combination with ACE2 blockage on SARS-CoV-2 infection (A) Schematic of the experimental owchart on the co-cultures. (B) The ACE2 blockage antibody was applied two hours prior to the virus presence, and the luciferase activity of the co-cultures of lung cells and M1, M2 macrophages (iMφ or THP-1) or 293T cells (control) was measured at Mock or infected with SARS-CoV-2 pseudo-entry virus at 24 hpi (MOI=0.01). P values were calculated by unpaired two-tailed Student's t test. ***P < 0.001, ****P < 0.0001. (C) The ACE2 blockage antibody was applied two hours prior to the virus presence, IF staining was performed on the co-cultures of iLung cells and iM1φ, iM2φ, or 293T, at Mock or infected with SARS-CoV-2 virus at 24 hpi (MOI=0.01), using antibodies detecting SARS-CoV-2 NSP14 protein, CD80 or CD206. ILung cells expressed GFP. Scale bar = 100 µm

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.