SARS-CoV-2 Infected Cardiomyocytes Recruit Monocytes by Secreting CCL2

Abstract Heart injury has been reported in up to 20% of COVID-19 patients, yet the cause of myocardial histopathology remains unknown. In order to study the cause of myocardial pathology in COVID-19 patients, we used a hamster model to determine whether following infection SARS-CoV-2, the causative agent of COVID-19, can be detected in heart tissues. Here, we clearly demonstrate that viral RNA and nucleocapsid protein is present in cardiomyocytes in the hearts of infected hamsters. Interestingly, functional cardiomyocyte associated gene expression was decreased in infected hamster hearts, corresponding to an increase in reactive oxygen species (ROS). This data using an animal model was further validated using autopsy heart samples of COVID-19 patients. Moreover, we show that both human pluripotent stem cell-derived cardiomyocytes (hPSC-derived CMs) and adult cardiomyocytes (CMs) can be infected by SARS-CoV-2 and that CCL2 is secreted upon SARS-CoV-2 infection, leading to monocyte recruitment. Increased CCL2 expression and macrophage infiltration was also observed in the hearts of infected hamsters. Using single cell RNA-seq, we also show that macrophages are able to decrease SARS-CoV-2 infection of CMs. Overall, our study provides direct evidence that SARS-CoV-2 infects CMs in vivo and proposes a mechanism of immune-cell infiltration and pathology in heart tissue of COVID-19 patients.


145
We further examined the response of adult human CMs to SARS-CoV-2 infection. Adult human 146 CMs were infected with SARS-CoV-2 (USA-WA1/2020, MOI=0.1). Similar to hPSC-derived 147 CMs, significant levels of viral subgenomic RNA (Fig. 2j) and robust read coverage across the 148 viral genome (Fig. 2k) were detected in adult human CMs. PCA and clustering analysis showed 149 that mock and SARS-CoV-2 infected adult human CM transcript profiles clustered separately ( Fig.   150 2l, 2m). Furthermore, consistent with data from hPSC-derived CMs, analysis of the host 151 transcriptional response revealed robust induction of chemokines, including CCL2 (Fig. 2n, 2o). 152 Consistent with hPSC-derived CMs (Fig. 2i), KEGG pathway analysis in adult human CMs 153 highlighted pathways involved in inflammatory and immune responses, including IL-17 signaling 154 pathway, TNF signaling pathway, cytokine-cytokine receptor interaction, and chemokine 155 signaling pathway (Fig. 2p). Finally, ELISA assays confirmed significantly increased levels of 156 CCL2 in the medium of H9 hESC-derived CMs after SARS-CoV-2 infection, compared to mock SARS-CoV-2 infected cardiomyocytes recruit monocytes by secreting CCL2. 170 Macrophages include tissue-resident macrophages and migrating macrophages 20 . Migrating 171 macrophages are typically derived from monocytes in the blood. During inflammation, circulating 172 monocytes leave the bloodstream and migrate into tissues where, following conditioning by local 173 growth factors, pro-inflammatory cytokines and microbial products, they differentiate into 174 macrophages 1 . CCL2 is a chemotactant for monocytes and basophils. As such, we hypothesized 175 that CCL2 expression of infected CMs attracts monocytes to the site of infection.

177
To investigate this hypothesis, we therefore examined the ability of SARS-CoV-2 infected CMs 178 to stimulate migration and recruitment of monocytes. Monocytes were derived from the same 179 parental H9 or H1 hESC line following a previously reported protocol 21 (Extended Data Fig. 2a) 180 through a stepwise manner, including the generation of mesodermal cells, followed by 181 hematopoietic progenitor cells, monocytes (Extended Data Fig. 2b), and finally CD14 + , CD11B + 182 macrophages (Extended Data Fig. 2c). To study recruitment, hPSC-derived CMs were plated on 183 the bottom of trans-well plates and hPSC-derived monocytes were plated on top of the insert (Fig.   184   4a). 24 hpi of CMs, the number of migrated monocytes was significantly higher when cultured with SARS-CoV-2 infected hPSC-derived CMs than when cultured with mock infected hPSC-186 derived CMs using two different hPSC-derive monocytes (Fig. 4b, 4c and Extended Data Fig.   187 3a, 3b). These findings were subsequently validated using adult human CMs, also showing that 188 monocytes were recruited at a significantly higher rate when cultured with infected rather than 189 mock infected adult human CMs (Fig. 4d, 4e and Extended Data Fig. 3c, 3d). 190 191 To determine whether CCL2 is sufficient to recruit monocyte, CCL2 was added to the lower level 192 of transwell plates with monocytes embedded in the insert. After 24 h after CCL2 treatment, a 193 significantly higher number of monocytes were found to have migrated to the bottom of the plate 194 compared to mock treated plates (Fig. 4f, 4g and Extended Data Fig. 3e, 3f). To determine 195 whether CCL2 is the key driver for monocyte migration, hPSC-derived or adult human CMs co-196 culture assays with monocytes were infected with SARS-CoV-2 in the presence or absence of 197 CCL2 neutralizing antibodies or a CCR2 inhibitor ( Fig. 4h-4k, and Extended Data Fig. 3g-3j). 198 When thereby blocking CCL2 action, significantly less migrating monocytes were detected after 199 viral infection. Together, these data suggest that monocytes are directly recruited to infected CMs Co-culture of hPSC-derived cardiomyocytes and macrophages reveals that macrophages can 203 reduce SARS-CoV-2 infection of CMs. 204 We next investigated how recruited macrophages affect the viral infection. To model the viral 205 entry process, we created an immunocardiac co-culture platform containing hPSC-derived CMs 206 and hPSC-derived macrophages. This immunocardiac co-culture was infected with SARS-CoV-2 207 entry virus carrying a luciferase (Luc) reporter (MOI=0.1) or mock-infected as described 208 previously 10 . At 24 hpi, cells were monitored for Luc activity. The presence of macrophages 209 significantly decreased the Luc activity in a dose-dependent manner (Extended Data Fig. 4a, 4b). 210 Immunostaining further confirmed the decrease of Luc + cells in MYH6:mCherry + cells (Extended 211 Data Fig. 4c, 4d). The immunocardiac co-culture was further examined by scRNA-seq at 24 hpi.

212
The transcript profiling data was projected using Uniform Manifold Approximation and Projection 213 (UMAP). In the virus-immunocardiac co-culture platform (immunocardiac co-culture infected 214 with virus), four distinct cell clusters were identified, including CMs, macrophages, 215 stem/progenitor cells, and one cluster expressing both CM and macrophage markers (Fig. 5a). The  The putative viral receptor ACE2 is expressed mainly in hPSC-derived CMs and cardiac 221 progenitors (Extended Data Fig. 4g, 4h). The effector protease TMPRSS2 22 is not obviously 222 expressed in hPSC-derived cardiac progenitors (Extended Data Fig. 4g, 4h). However, FURIN, 223 the gene encoding a pro-protein convertase that pre-activates SARS-CoV-2 23 , and CTSL, the gene 224 encoding cathepsin L a proteinase that might be able to substitute for TMPRSS2 22 , are highly 225 expressed in both hPSC-derived CMs and cardiac progenitors (Extended Data Fig. 4g, 4h).

227
The mRNAs derived from SARS-CoV-2 entry virus, including Luc, were detected in infected CMs, 228 but at very low levels in macrophages (Extended Data Fig. 4i, 4j), which is consistent with our 12 previous report 24 . The one cell cluster that expressed markers of both CMs and macrophages, and 230 in addition high levels of viral genes, likely represents infected CMs engulfed by macrophages 231 (Fig. 5c). The Luc expression in CMs of virus-immunocardiac co-cultures was much lower than 232 that of virus infected CMs (Fig. 5d, 5e), suggesting that macrophages decreased the infection of 233 SARS-CoV-2-pseudo entry virus to CMs. Consistently, the infected CMs show increased 234 expression of CCL2 (Extended Data Fig. 4k, 4l).

236
To further validate the impact of macrophages on SARS-CoV-2 infection, the immunocardiac co-237 culture platform containing hPSC-derived CMs and hPSC-derived macrophages were infected 238 with SARS-CoV-2 (MOI=0.1) or mock infected. At 24 hpi, cells were analyzed using either qRT-239 PCR or immunostaining. The qRT-PCR of replicating viral RNA normalized to a cardiomyocyte 240 marker, cTNT, suggested significantly decreased SARS-CoV-2 infection (Fig. 5f). 241 Immunostaining further validated the decrease of SARS-CoV-2 + in cTNT + cells (Fig. 5g, 5h and 242 Extended Data Fig. 4m, 4n). We further performed long-term co-culture of hPSC-derived CMs 243 and macrophages and confirmed that the presence of macrophages decreased SARS-CoV-2 244 infection to CMs when co-cultured with macrophages for one week (Fig. 5i, 5j). Together, the 245 data suggest that macrophages decrease SARS-CoV-2 infection of CMs.   (Fig. 1d), suggesting that macrophages recruited by CMs might also contribute to      To differentiate cardiomyocytes (CMs) from hPSC, hPSCs were passaged at a density of 325 3x10 5 cells/well of 6-well plate and grown for 48 hours in a humidified incubator with 5% CO2 at 326 37℃ to reach 90% confluence. On day 0, the medium was replaced with RPMI 1640 supplemented 327 with B27 minnus insulin and 6 µM CHIR99021. On day 1, the medium was changed to RPMI   The differentiation protocol was adapted from a previously reported protocol 21 . First, hPSC cells 348 were lifted with ReLeSR (STEMCELL Technologies) as small clusters onto Matrigel-coated 6-349 well plates at a low density. After 1 day, medium was refreshed with IF9S medium supplemented 350 with 50 ng/ml BMP-4, 15 ng/ml Activin A and 1.5 µm CHIR99021. On day 2, medium was 351 refreshed with IF9S medium supplemented with 50 ng/ml VEGF, 50 ng/ml bFGF, 50 ng/ml SCF 352 (R&D Systems) and 10 µm SB431542 (Cayman Chemical). On day 5 and day7, medium was 353 changed into IF9S supplemented with 50 ng/ml IL-6 (R&D Systems), 12 ng/ml IL-3 (R&D 354 Systems), 50 ng/ml VEGF, 50 ng/ml bFGF, 50 ng/ml SCF and 50 ng/ml TPO (R&D Systems). On The migration of macrophages was examined using 24 well Trans-well chambers (6.5 mm insert;  The immunocardiac co-culture 370 hPSC-derived cardiomyocytes were dissociated with Accutase for 5-10min at 37℃ followed by 371 resuspending with fresh RPMI 1640 plus normal B27 and Y-27632 and reseeding into plates. After 372 24 h recovery, the medium was switched to RMPI 1640 plus B27 without Y-27632. After another 373 24 h recovery, hPSC-derived macrophages were dissociated with Accutase for 3 min and added 374 into hPSC-derived cardiomyocytes. The immunocardiac co-culture cells were cultured for another 375 24 h (short-term co-culture) or 7 days (long-term co-culture) before following analysis. Adult 376 cardiomyocytes were also seeded into plates for 48-96 h and co-cultured with hPSC-derived 377 macrophages for another 24 h before following analysis.   Single-cell RNA-seq data analysis 436 We filtered a small fraction of cells with viral gene content greater than 80% but less than 200 437 genes detected for which we believe are not real cells but rather empty beads with ambient RNAs. 438 We then filtered cells with less than 400 or more than 7000 genes detected as well as cells with were generated using the R ggplot2 package.      ATP1B1  RYR2  ATP2A2  SLC8A1  KCNH2  ATP2B4  SORBS2  ACTN2  NEBL  PLN  COX17  NDUFA9  MYL2  ACTA2  MYL7  ACTN2  MYH7  MYL3  TNNC1  SORBS1  ALPK3  IRX4  MEF2A  TBX5  GATA4  MEF2C  PPARGC1A  MEF2D  MYOCD   CALR  GPX7  DNAJB2  MPV17  PREX1  ATOX1  CALR  CCS  CCT2  DNAJB1  DNAJC7  DNAJC8  FTH1  HMOX1  MPV17  PRDX1  PRDX4     CMs infected with SARS-CoV-2 virus or mock (MOI=0.1). N=3 independent biological replicates.

563
Data was presented as mean ± STDEV. P values were calculated by unpaired two-tailed Student's     were calculated by unpaired two-tailed Student's t test. *P < 0.05, **P < 0.01, and ***P < 0.001.  Immunostaining (i) and quantification (j) of hPSC-derived CMs at 24 hpi with mock or SARS-608 CoV-2 in the presence or absence of macrophages (MOI=0.1) for long-time co-culture (7 days).