key: cord-0283019-z4oznffj
authors: Navaratnarajah, Chanakha K.; Pease, David R.; Halfmann, Peter; Taye, Biruhalem; Barkhymer, Alison; Howell, Kyle G.; Charlesworth, Jon E.; Christensen, Trace A.; Kawaoka, Yoshihiro; Cattaneo, Roberto; Schneider, Jay W.
title: Highly efficient SARS-CoV-2 infection of human cardiomyocytes: spike protein-mediated cell fusion and its inhibition
date: 2021-07-30
journal: bioRxiv
DOI: 10.1101/2021.07.30.454437
sha: b972b61c42aa888935f71a424f45de43e509c54f
doc_id: 283019
cord_uid: z4oznffj

Severe cardiovascular complications can occur in coronavirus disease of 2019 (COVID-19) patients. Cardiac damage is attributed mostly to a bystander effect: the aberrant host response to acute respiratory infection. However, direct infection of cardiac tissue by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also occurs. We examined here the cardiac tropism of SARS-CoV-2 in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM) that beat spontaneously. These cardiomyocytes express the angiotensin I converting-enzyme 2 (ACE2) receptor and a subset of the proteases that mediate spike protein cleavage in the lungs, but not transmembrane protease serine 2 (TMPRSS2). Nevertheless, SARS-CoV-2 infection was productive: viral transcripts accounted for about 88% of total mRNA. In the cytoplasm of infected hiPSC-CM, smooth walled exocytic vesicles contained numerous 65-90 nm particles with typical ribonucleocapsid structures, and virus-like particles with knob-like spikes covered the cell surface. To better understand the mechanisms of SARS-CoV-2 spread in hiPSC-CM we engineered an expression vector coding for the spike protein with a monomeric emerald-green fluorescent protein fused to its cytoplasmic tail (S-mEm). Proteolytic processing of S-mEm and the parental spike were equivalent. Live cell imaging tracked spread of S-mEm signal from cell to cell and documented formation of syncytia. A cell-permeable, peptide-based molecule that blocks the catalytic site of furin abolished cell fusion. A spike mutant with the single amino acid change R682S that inactivates the furin cleavage site was fusion inactive. Thus, SARS-CoV-2 can replicate efficiently in hiPSC-CM and furin activation of its spike protein is required for fusion-based cytopathology. This hiPSC-CM platform provides an opportunity for target-based drug discovery in cardiac COVID-19. Author Summary It is unclear whether the cardiac complications frequently observed in COVID-19 patients are due exclusively to systemic inflammation and thrombosis. Viral replication has occasionally been confirmed in cardiac tissue, but rigorous analyses are restricted to rare autopsy materials. Moreover, there are few animal models to study cardiovascular complications of coronavirus infections. To overcome these limitations, we developed an in vitro model of SARS-CoV-2 spread in induced pluripotent stem cell-derived cardiomyocytes. In these cells, infection is highly productive: viral transcription levels exceed those documented in permissive transformed cell lines. To better understand the mechanisms of SARS-CoV-2 spread we expressed a fluorescent version of its spike protein that allowed to characterize a fusion-based cytopathic effect. A mutant of the spike protein with a single amino acid mutation in the furin cleavage site lost cytopathic function. The spike protein of the Middle East Respiratory Syndrome (MERS) coronavirus drove cardiomyocyte fusion with slow kinetics, whereas the spike proteins of SARS-CoV and the respiratory coronavirus 229E were inactive. These fusion activities correlated with the level of cardiovascular complications observed in infections with the respective viruses. These data indicate that SARS-CoV-2 has the potential to cause cardiac damage by fusing cardiomyocytes.

lethal systemic symptoms [1] . 

Expression of virus entry factors in cardiomyocytes 94 We assessed whether ACE2, the SARS-CoV-2 receptor, and the spike-activating proteases 95 TMPRSS2 and cathepsin B (CTSB) are expressed during the differentiation process of human 96 embryonic stem cells into cardiomyocytes. The ACE2 transcription level peaked at day-20, those 97 of cathepsin B remained stable, and TMPRSS2 transcripts were never detectable (S1 Table) . 98

Thus, we characterized our hiPSC-CM at this differentiation stage. Super resolution 99

immunofluorescence confocal microscopy documented cell surface expression of ACE2, and the 100 striated F-actin organization typical of cardiomyocytes (Fig. 1A) . In particular, ACE2 receptors 101 clustered in raft-like puncta diffusely distributed across the sarcolemma and extended into 102 filopodia contacting adjacent cardiomyocytes (Fig. 1A, arrow highlights filopodia) . 103

To further characterize day-20 differentiated cardiomyocytes, we analyzed their total cellular 104 transcriptome by RNAseq. Cardiomyocyte differentiation markers were expressed in our hiPSC-CM at levels similar to those documented in two other hiPSC-CM lines used for SARS-CoV-2 106 infection studies (Fig. 1B , compare N with P and S panels). The ACE2 receptor was expressed at 107 higher levels in our cardiomyocytes than in those used in the other studies. In all three studies 108 transcripts of the proteases cathepsins B, cathepsin L, and furin, were detected, but transcripts 109 of the protease (TMPRSS2) that enables endosome independent viral entry in the lungs [24] , 110 were below detection levels (less than 0.5 counts/million in at least 2 samples). 111

Highly productive cardiomyocyte infection 112 We inoculated two independent lines of spontaneously beating hiPSC-CMs with SARS-CoV-2 at 113 0.01 multiplicity of infection (MOI) and monitored virus titer in the supernatant by plaque 114 assay. Two days after inoculation about 10 6 infectious units/ml were produced ( Fig. 2A) . We then sought to document expression, processing and localization of the viral proteins. 

We also monitored the cytopathic effects of SARS-CoV-2 infection of hiPSC-CMs by IF confocal 149 microscopy. In Fig. 5A the nuclei of infected cells were stained with DAPI (blue), and the viral M 150 protein and cytoskeletal alpha-actinin with specific antibodies (green and red, respectively). To characterize the mechanism of cell fusion, we engineered a SARS-CoV-2 full-length 165 recombinant spike protein fused to modified Emerald green fluorescent protein at its carboxyl-166 terminus (CoV-2 S-mEm) (Fig. 6A , left panel). We validated this reagent in Vero cells that, like 167 hiPSC-CMs, express ACE2 but not TMPRSS2. In these cells CoV-2 S-mEm was appropriately cleaved (Fig. 6A , right panel). Super resolution confocal microscopy localized CoV-2 S-mEm to 169 hair-like plasma membrane extensions (Fig. 6B) . Fluorescent activated cell sorting confirmed 170

CoV-2 S-mEm cell surface expression (Fig. 6C) . Live cell imaging tracked spread of CoV-2 S-mEm 171 signal from cell to cell through membrane fusion, generating syncytia ( Fig. 6D and S2 Movie). 172

We then assessed whether CoV-2 S-mEm fuses cardiomyocytes. Despite overall transfection 173 efficiency <5%, CoV-2 S-mEM expressing hiPSC-CMs produced syncytia with nuclei frequently 174 arranged in clusters or rosettes (Fig. 7A , and S3 Movie). Some syncytia were characterized by 175 circular or oval enucleated cytoskeletal "corpses" shown by F-actin phalloidin staining ( Knowing that furin, a protease located in the trans-Golgi apparatus that contributes to SARS-182

CoV-2 spike activation, is expressed in hiPSC-CM (Fig. 1B) , we sought to block its function 183 biochemically and genetically. For biochemical interference we used Decanoyl-RVKR-CMK (furin 184 inhibitor, FI), a cell-permeable peptide-based molecule that irreversibly blocks its catalytic site. 185

For genetic interference, we generated an expression vector differing from CoV-2 S through the We validated these approaches in Vero cells. The left panel of Fig. 8A documents progressive 189 inhibition of CoV-2 S protein processing (S0 cleavage into S1 and S2) by increasing 190 concentrations of FI. The second and third panels show that fusion occurs in cells expressing 191

CoV-2 S in the absence of FI, but not in its presence. The last panel shows that the R682S 192 Bioinformatics and data analysis 304

The quality of the raw RNA-seq data was assessed by fastqp v0.20.1 [44] , and quality reads 305 were filtered and aligned against human genome (hg19) using STAR alignment (v2.7.8a) [45] in 306 galaxy platform (https://usegalaxy.org). The aligned reads were counted using htseq-count 307 v0.9.1 [46] and 0.5 read counts per million (CPM) in at least two samples was used as an 308 expression threshold. Trimmed mean of M values normalized (TMM) [47] and log2 transformed 309 data was used for plotting heatmaps and differential analysis in limma [48] . For the detection of 

Coronaviridae: The Viruses and Their Replication

Fields Virology: Emerging Viruses

Cardiovascular Implications of Fatal 449 Outcomes of Patients With Coronavirus Disease 2019 (COVID-19)

COVID-19 can affect the heart

Potential Effects of Coronaviruses on 455 the Cardiovascular System: A Review

COVID-458 19 and Cardiovascular Disease: From Bench to Bedside

COVID-19, and the Renin-Angiotensin System

Pressing Needs and Best Research Practices

Characteristics of and Important Lessons From the Coronavirus 467 Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From

Cardiac Infection With SARS-CoV-2 in Confirmed COVID-19 Autopsy Cases

SARS-CoV-2 genomes in endomyocardial biopsy with latency after acute infection

Epub 2020/10/13

Pathogenesis of SARS-CoV-2 induced cardiac injury from the perspective of 499 the virus

Myocarditis is rare in COVID-19 autopsies: cardiovascular 503 findings across 277 postmortem examinations

Pathological 507 Evidence for SARS-CoV-2 as a Cause of Myocarditis: JACC Review Topic of the Week

CoV-2 Reverse Genetics Reveals a Variable Infection Gradient in the Respiratory Tract

Morphogenesis and cytopathic effect 583 of SARS-CoV-2 infection in human airway epithelial cells

Syncytia formation 587 by SARS-CoV-2-infected cells

A Multibasic Cleavage Site in the Spike Protein 591 of SARS-CoV-2 Is Essential for Infection of Human Lung Cells

Drugs that inhibit 595 TMEM16 proteins block SARS-CoV-2 spike-induced syncytia

Role of host factors in SARS-CoV-2 entry

Furin cleavage 601 of SARS-CoV-2 Spike promotes but is not essential for infection and cell-cell fusion

Ready, set, fuse! The coronavirus spike protein and 605 acquisition of fusion competence

Avoiding culture shock with the SARS-CoV-2 spike protein

fastp: an ultra-fast all-in-one FASTQ preprocessor

STAR: ultrafast 615 universal RNA-seq aligner

HTSeq--a Python framework to work with high-throughput 619 sequencing data

A scaling normalization method for differential expression 623 analysis of RNA-seq data

limma powers differential 627 expression analyses for RNA-sequencing and microarray studies

Alignment/Map format and SAMtools

Histologic fixatives suitable for diagnostic light and electron 635 microscopy

Measles viruses on throat 638 swabs from measles patients use signaling lymphocytic activation molecule

not CD46 as a cellular receptor

Preferential initiation at the second 643 AUG of the measles virus F mRNA: a role for the long untranslated region

Neuropilin-1 facilitates SARS-CoV-2 cell entry and infectivity

Neuropilin-1 is a host factor for SARS-CoV-2 infection

A) IF super resolution confocal microscopy 657 analysis of ACE2 and F-actin expression in hiPSC-CMs. Scale bar, 10 µm. (B) Transcript levels of 658 cardiomyocytes marker genes and of virus entry factors. Scale on the right: log2 trimmed mean 659 of M values (TMM) normalized sequence read counts. N=this study

Gene abbreviations: FHL2=four and a half LIM domains

MYBPC3=myosin binding protein C

NPPA=natriuretic peptide A; PLN=phospholamban; RYR2=ryanodine receptor 2

TNNT2=troponin T type 2; TTN=titin

SARS-CoV-2 titers in two hiPSC-CM cell 670 lines: open squares, hiPSC-CM#1; filled dots, hiPSC-CM#2, each data point represents one 671 biological replicate. (B) Quantification of viral transcripts in infected hiPSC-CMs from this study 672 (N) and two published studies

C-E) Companion immunoblots 674 (left) and low-power IF confocal microscopy (right) of (C) SARS-CoV-2 spike glycoprotein

D) nucleocapsid, (N) and (E) membrane (M) protein, monomer (m) and insoluble aggregate 676 (a) in hiPSC-CMs, 48 hours post-infection. Scale bar

Cytopathic effects of SARS-CoV-2 in hiPSC-CMs. (A-B) IF confocal microscopy of SARS

C-D) IF super resolution confocal microscopy of infected and mock-infected hiPSC-CMs, 697 respectively. Scale bars, 10 µm. (E) Quantification of cell fusion in SARS-CoV-2 infected and 698 mock infected hiPSC-CMs 48 hours post-inoculation. Polyploidy index is the average number of 699 syncytia per field (n= 3 biological replicates). Nuclearity index is the average number of nuclei 700 per cell per field (n= 3 biological replicates, p = 0.0002, two-tailed t-test). Box and whisker plots 701 show median

Expression and function of SARS CoV-2 spike protein tagged with mEmerald. (A) Left 706 panel: schematic of SARS-CoV-2 S tagged with mEmerald (mEm) at the cytoplasmic tail

Cleavage at the S1/S2 furin site primes the spike protein for activation. S1, S1 subunit; S2, S2 708 subunit; N-/C-RBD, N-/C-terminal receptor binding domains

TM, trans-membrane segment

The fusion peptide is shown in blue and heptad repeat 1 and 2 in magenta and dark magenta, 710 respectively. The location of the furin cleavage mutant, R682S is indicated. The monoclonal antibody 1A9, which was used to detect the spike proteins, binds to an exposed loop (purple) 712 located close to heptad repeat 2. Right panel: immunoblot of the CoV-2 S and CoV-2 S-mEm 713 proteins detecting their S0 and S2 subunits

Cellular localization of the tagged 715 spike protein in non-permeabilized HeLa cells transfected with the expression plasmid for S-716 mEm. This protein was detected either by fluorescence emission (horizontal axis) or by using 717 spike-specific-mAb 1A9 and AF647 conjugated secondary-antibody

Vero cells transfected with CoV-2 S-mEm are indicated by a dotted yellow line. Scale bar

SARS-CoV-2 spike protein induces syncytia in hiPSC-CMs. (A) IF confocal microscopy of

SARS-CoV-2 spike (CoV-2 S)-expressing hiPSC-CMs. Scale bar, 50 µm. Viral and cellular 724 components visualized are indicated with their corresponding color below each panel

confocal microscopy of CoV-2 S-expressing hiPSC-CM with enucleated actin cytoskeletal 726 "corpses" (yellow arrows). (C) Super resolution confocal microscopy of CoV-2 S-mEM 727 localization to hiPSC-CM filopodia directly contacting the sarcolemma of an adjacent hiPSC-CM 728 (yellow circle)

A) (left panel) immunoblot analysis of CoV-2 S protein processing (S0 733 cleavage into S1 and S2) in Vero cells treated with increasing concentrations of FI (0 μM, 5 μM, 734 10 μM and 20 μM); cell lysates were separated by 4-15% SDS-PAGE under reducing conditions. 735 (2 nd to 4 th panels) phase contrast images of Vero cells expressing CoV-2 S in the absence or 736 presence of 20 µM FI, or of Vero cells expressing the R682S cleavage mutant, respectively, 72-737 hours after transfection

nd to 4 th panels) 741 phase contrast images of hiPSC-CM expressing CoV-2 S in the absence or presence of 20 µM FI, 742 or of Vero cells expressing the R682S cleavage mutant, respectively, 72-hours after 743 transfection

Box and 746 whisker plots for all quantification in this figure shows median, upper and lower quartile, and 747 extremes. 748 protein-mediated hiPSC syncytia, largest example circled in red. Scale bar, 50 µm. (C) Bright 755 field microscopy of crystal violet-stained hiPSC-CM expressing recombinant MERS spike protein 756 at 5 days post-transfection

Intercellular spread of CoV-2 S-mEm spike protein and development of Vero cell 762 syncytia

Intercellular spread of CoV-2 S-mEm spike protein and development of hiPSC-CM 764 syncytia

ACE2 and TMPRSS2 expression in H9 human embryonic stem cells

* Affymetrix microarray numerical values across an individual probe set

present): transcript is significantly (P <0.05) expressed compared with perfectly matched 769 and mismatched (background) probe sets