Efferocytosis of SARS-CoV-2-infected dying cells impairs macrophage anti-inflammatory functions and clearance of apoptotic cells

  1. Ana CG Salina
  2. Douglas dos-Santos
  3. Tamara S Rodrigues
  4. Marlon Fortes-Rocha
  5. Edismauro G Freitas-Filho
  6. Daniel L Alzamora-Terrel
  7. Icaro MS Castro
  8. Thais FC Fraga da Silva
  9. Mikhael HF de Lima
  10. Daniele C Nascimento
  11. Camila M Silva
  12. Juliana E Toller-Kawahisa
  13. Amanda Becerra
  14. Samuel Oliveira
  15. Diego B Caetité
  16. Leticia Almeida
  17. Adriene Y Ishimoto
  18. Thais M Lima
  19. Ronaldo B Martins
  20. Flavio Veras
  21. Natália B do Amaral
  22. Marcela C Giannini
  23. Letícia P Bonjorno
  24. Maria IF Lopes
  25. Maira N Benatti
  26. Sabrina S Batah
  27. Rodrigo C Santana
  28. Fernando C Vilar
  29. Maria A Martins
  30. Rodrigo L Assad
  31. Sergio CL de Almeida
  32. Fabiola R de Oliveira
  33. Eurico Arruda Neto
  34. Thiago M Cunha
  35. José C Alves-Filho
  36. Vania LD Bonato
  37. Fernando Q Cunha
  38. Alexandre T Fabro
  39. Helder I Nakaya
  40. Dario S Zamboni
  41. Paulo Louzada-Junior
  42. Rene DR Oliveira
  43. Larissa D Cunha

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    In this paper, the authors demonstrate that the macrophage efferocytic response in SARS-CoV-2 infection is compromised for two reasons. Internalization of apoptotic SARS-CoV-2 infected cells leads to: 1) A proinflammatory as opposed to an anti-inflammatory response; and 2) reduction in macrophage capacity to perform further efferocytosis

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

COVID-19 is a disease of dysfunctional immune responses, but the mechanisms triggering immunopathogenesis are not established. The functional plasticity of macrophages allows this cell type to promote pathogen elimination and inflammation or suppress inflammation and promote tissue remodeling and injury repair. During an infection, the clearance of dead and dying cells, a process named efferocytosis, can modulate the interplay between these contrasting functions. Here, we show that engulfment of SARS-CoV-2-infected apoptotic cells exacerbates inflammatory cytokine production, inhibits the expression of efferocytic receptors, and impairs continual efferocytosis by macrophages. We also provide evidence supporting that lung monocytes and macrophages from severe COVID-19 patients have compromised efferocytic capacity. Our findings reveal that dysfunctional efferocytosis of SARS-CoV-2-infected cell corpses suppresses macrophage anti-inflammation and efficient tissue repair programs and provides mechanistic insights for the excessive production of pro-inflammatory cytokines and accumulation of tissue damage associated with COVID-19 immunopathogenesis.

Article activity feed

  1. Version published to 10.7554/elife.74443 on eLife
  2. Version published to 10.1101/2021.02.18.21251504v3 on medRxiv
  3. eLife

    Author Response

    Reviewer #1 (Public Review):

    Salina, dos-Santos, Rodrigues, et al demonstrate that dying cells from SARS-CoV-2infected cultures shift gene expression from an alternative activation-like state (characterized by CD206) towards a classical activation-like state (characterized by IL6) for primary human macrophages and the THP-1 macrophage cell line. These phenotypes of reduced CD206 expression and elevated IL6 expression were not induced by SARS-CoV-2 "AC" (loosely adherent cells in culture) after UV sterilization, annexin V treatment, cytochalasin treatment (to inhibit internalization), or fixation. These phenotypes were also not replicated by supernatant factors from SARS-CoV-2 infected cultures. Furthermore, coxsackievirus-infected dying cells did not induce similar effects on macrophages as SARS-CoV-2-infected dying cells. Uptake of SARS-CoV-2 ACs led to reduced macrophage expression of phosphatidylserine (PtSer) receptors and reduced uptake of more apoptotic cells. Upon autopsy of deceased COVID-19 patients, the authors found reduced CD36 and MERTK on lung phagocytes by microscopy, consistent with their in vitro findings. Furthermore, reanalyzed published scRNA-Seq data from broncheoalveolar lavage indicated that expression of several efferocytosisrelated modules was decreased, especially in cells with SARS-CoV-2 mRNA. They also find that circulating monocytes from patient blood are altered in composition and show similar alterations. When circulating PMBC gene transcription was assessed by qPCR, they found similar reductions, which were not replicated in acute respiratory distress syndrome patients. This signature was correlated with more severe disease. Patient monocytes had specific defects in dead cell uptake. Taken together, Salina, dosSantos, Rodrigues et al demonstrate that SARS-CoV-2-infected dying cells induce changes in efferocytosis that are dependent on live virus, internalization of the dying cell, and PtSer recognition. While the authors describe and characterize several true (and intriguing phenomenon), with careful use of controls, I have one major concern, as well as several other concerns with the manuscript as currently constructed.

    We thank the reviewer for the detailed description and positive assessment of our work, and hope that the additions to the paper (our extensive experiments to address other points) will meet with approval.

    Major:

    1. The authors do not characterize the level of necrotic cells in their SARS-CoV-2 infected cultures. These will be present in their loosely adherent "AC" fraction and would be far more likely to induce IL-6. The authors never show costaining for Annexin V and a membrane impermeable dye (such as 7-AAD). This is a major oversight which must be addressed, as annexin V will stain both apoptotic and necrotic cells (in the first case, because PtSer is flipped. In the second case, because membrane integrity is lost). While their cleaved caspase 3 staining and use of zVAD is nice to address apoptosis more selectively, the annexin V staining as used is not sufficient. Most importantly, for their stimulation experiments, the authors need to find a way to separate the necrotic and apoptotic cell fractions, or otherwise address the role of necrotic cells. Otherwise, their findings could be due to necrotic cells (as would be more consistent with the considerable proinflammatory effects of the SARS-CoV-2 "AC" fraction).

    To address the reviewer’s concerns, we first performed co-staining of cells with Annexin V and a cell viability dye and evaluated them by flow cytometry. Our results show that at 48h post-infection, most dying cells are apoptotic (Ann+ Zombie-), similar to UV-irradiate cells (Fig. 1C). To further probe the possible occurrence of necrosis (or non-apoptotic regulated cell death) in response to SARS-CoV-2, we also performed a cytotoxicity assay. We confirmed that infection induced did not induce robust release of LDH, which would be expected in permeabilized cells (Fig. 1D). We also highlight that our former data assessing apoptosis by flow cytometry of active caspase-3 (Fig. 1B) has been complemented with immunoblot detection of cleaved caspase-3 and caspase-8 in infected cells (Fig. 1A), which we believe unequivocally supports the occurrence of apoptosis.

    Even if we identified few permeabilized annexin V+ dead cells in response to SARSCoV-2 at the evaluated time points (Fig. 1C), we strived to confirm that the presence of these cells in our isolated AC were not responsible for macrophage function modulation. In Fig. 3 – figure supplement 3D, we show that IL-6 secretion in response to infected AC still occurs when infected epithelial cells are cultivated in the presence of glycine, previously show to inhibit secretion of cell content by membrane permeabilization during necrotic or pyroptotic cell death. Further, we now provide data where macrophages were stimulated with intact versus permeabilized infected Ann V+ obtained by cell sorting. Our results demonstrate that IL-6 secretion occurs only in response to stimulation with apoptotic, non-permeabilized infected apoptotic cells (Fig. 3C). Collectively, these results support that the release of proinflammatory mediators or phagocytosis of necrotic cells, even if those occur in a small fraction, do not account for activation of macrophages.

    Minor

    1. As acknowledged by the authors, there is a major disconnect between their in vitro data and their patient data. As the authors clearly and elegantly demonstrate, soluble factors from SARS-CoV-2 infected cultures are inadequate to show many of the described affects of SARS-CoV-2 AC. Yet, the patient comparison done by the authors is with circulating (i.e., non-SARS-CoV-2 exposed) PBMC. While I appreciate that the authors are limited in the cell types they can obtain from SARS-CoV-2 infected patients, it is nonetheless a significant issue that the in vitro and (some) ex vivo portions of their study seemingly describe entirely different phenomenon.

    We appreciate the thoroughness of the reviewer’s assessment. We believe that our findings in patient monocytes are important as they suggest a possible broad impact on the efferocytic capacity of professional phagocytes during COVID-19, and presenting them even without a clear mechanistic insight, would be warranted in such a unique scenario as the pandemic. It remains possible that mediators associated with immune dysfunction and cells other than macrophages contribute to signal efferocytosis suppression systemically. However, we consider that understanding the interplay between local and systemic phenomena will take a considerable amount of work that surpasses the scope of this revision process, and were not able to perform a thorough investigation during the review process to establish a mechanistic link between in vitro and patient findings. Therefore, we decided to follow the reviewer’s and editorial suggestion and limit the data presented in the current manuscript to those concerning a direct effect of efferocytosis of infected apoptotic cells. We are very interested in understanding how SARS-CoV-2 infection in the lungs affects efferocytic capacity systemically and expect to publish our findings in a follow-up study.

    1. Figure 1B. Why do the staining patterns of spike and CD68 look identical? What controls do the authors have to detect and compensate for spillover between channels? Please explain this anomaly and how images were processed post-acquisition (software, etc).

    To rule out that spillover due to the chosen fluorophores (especially conjugation of Spike to low intensity FITC fluorophore), we repeated our experiment changing fluorophore combination (Fig. 1F). While we did see improvement of Spike signal in our new experimental setting, we still observe that CD68 and Spike staining superimpose at some, but not all, sites. As CD68 is described to be internalized in endosomes, it is possible that CD68 is also internalized with viruses and infected cells (as suggested by the “a” inset in Fig. 1F). Lung tissue of a COVID-19 patient incubated with secondary antibodies only and Alexa-568-conjugated Spike, were used as control to confirm that our acquisitions settings were not resulting in spillover. Images were acquired with an Eclipse Ti2-E microscope using the NIS-Elements acquisition software (Nikon). Brightness and contrast levels were adjusted in ImageJ image processing package (NIH). Image layouts were built on Adobe Photoshop. This information is now stated on Material and Methods.

    Reviewer #2 (Public Review):

    This study defined the cellular mechanisms of macrophages in severe SARS-CoV-2 infection. Using patients' samples and cell culture experiments, they demonstrated that SARS-CoV-2 switched macrophages from anti-inflammatory to pro-inflammatory phenotypes. The process of clearing apoptotic cells by macrophages was impaired in severe SARS-CoV-2 infection. The macrophages accumulated the dying cells inside excessively expressed inflammatory genes. The study is significant, indicating the potential molecular targets to ameliorate severe SARS-CoV-2 infection. The logical demonstration that "sensing and engulfment of dying cells carrying viable SARS-CoV-2" (line 198) but not other pathogens switched macrophages toward the pro-inflammatory phenotypes is clear. The manuscript will be more improved if the authors test the impact of COVID19 pills and vaccines on their phenotypes in efferocytosis.

    We are pleased that the reviewer felt that this study could be of interest to the eLife readership, and have endeavored to improve the manuscript following some of the reviewer’s suggestions.

  4. eLife

    Evaluation Summary:

    In this paper, the authors demonstrate that the macrophage efferocytic response in SARS-CoV-2 infection is compromised for two reasons. Internalization of apoptotic SARS-CoV-2 infected cells leads to: 1) A proinflammatory as opposed to an anti-inflammatory response; and 2) reduction in macrophage capacity to perform further efferocytosis

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.)

  5. eLife

    Reviewer #1 (Public Review):

    Salina, dos-Santos, Rodrigues, et al demonstrate that dying cells from SARS-CoV-2-infected cultures shift gene expression from an alternative activation-like state (characterized by CD206) towards a classical activation-like state (characterized by IL6) for primary human macrophages and the THP-1 macrophage cell line. These phenotypes of reduced CD206 expression and elevated IL6 expression were not induced by SARS-CoV-2 "AC" (loosely adherent cells in culture) after UV sterilization, annexin V treatment, cytochalasin treatment (to inhibit internalization), or fixation. These phenotypes were also not replicated by supernatant factors from SARS-CoV-2 infected cultures. Furthermore, coxsackievirus-infected dying cells did not induce similar effects on macrophages as SARS-CoV-2-infected dying cells. Uptake of SARS-CoV-2 ACs led to reduced macrophage expression of phosphatidylserine (PtSer) receptors and reduced uptake of more apoptotic cells. Upon autopsy of deceased COVID-19 patients, the authors found reduced CD36 and MERTK on lung phagocytes by microscopy, consistent with their in vitro findings. Furthermore, reanalyzed published scRNA-Seq data from broncheoalveolar lavage indicated that expression of several efferocytosis-related modules was decreased, especially in cells with SARS-CoV-2 mRNA. They also find that circulating monocytes from patient blood are altered in composition and show similar alterations. When circulating PMBC gene transcription was assessed by qPCR, they found similar reductions, which were not replicated in acute respiratory distress syndrome patients. This signature was correlated with more severe disease. Patient monocytes had specific defects in dead cell uptake. Taken together, Salina, dos-Santos, Rodrigues et al demonstrate that SARS-CoV-2-infected dying cells induce changes in efferocytosis that are dependent on live virus, internalization of the dying cell, and PtSer recognition. While the authors describe and characterize several true (and intriguing phenomenon), with careful use of controls, I have one major concern, as well as several other concerns with the manuscript as currently constructed.

    Major:

    1. The authors do not characterize the level of necrotic cells in their SARS-CoV-2 infected cultures. These will be present in their loosely adherent "AC" fraction and would be far more likely to induce IL-6. The authors never show costaining for Annexin V and a membrane impermeable dye (such as 7-AAD). This is a major oversight which must be addressed, as annexin V will stain both apoptotic and necrotic cells (in the first case, because PtSer is flipped. In the second case, because membrane integrity is lost). While their cleaved caspase 3 staining and use of zVAD is nice to address apoptosis more selectively, the annexin V staining as used is not sufficient. Most importantly, for their stimulation experiments, the authors need to find a way to separate the necrotic and apoptotic cell fractions, or otherwise address the role of necrotic cells. Otherwise, their findings could be due to necrotic cells (as would be more consistent with the considerable proinflammatory effects of the SARS-CoV-2 "AC" fraction).
      Minor
    2. As acknowledged by the authors, there is a major disconnect between their in vitro data and their patient data. As the authors clearly and elegantly demonstrate, soluble factors from SARS-CoV-2 infected cultures are inadequate to show many of the described affects of SARS-CoV-2 AC. Yet, the patient comparison done by the authors is with circulating (i.e., non SARS-CoV-2 exposed) PBMC. While I appreciate that the authors are limited in the cell types they can obtain from SARS-CoV-2 infected patients, it is nonetheless a significant issue that the in vitro and (some) ex vivo portions of their study seemingly describe entirely different phenomenon.
    3. Figure 1B. Why do the staining patterns of spike and CD68 look identical? What controls do the authors have to detect and compensate for spillover between channels? Please explain this anomaly and how images were processed post-acquisition (software, etc).
  6. eLife

    Reviewer #2 (Public Review):

    This study defined the cellular mechanisms of macrophages in severe SARS-CoV-2 infection. Using patients' samples and cell culture experiments, they demonstrated that SARS-CoV-2 switched macrophages from anti-inflammatory to pro-inflammatory phenotypes. The process of clearing apoptotic cells by macrophages was impaired in severe SARS-CoV-2 infection. The macrophages accumulated the dying cells inside excessively expressed inflammatory genes. The study is significant, indicating the potential molecular targets to ameliorate severe SARS-CoV-2 infection. The logical demonstration that "sensing and engulfment of dying cells carrying viable SARS-CoV-2" (line 198) but not other pathogens switched macrophages toward the pro-inflammatory phenotypes is clear. The manuscript will be more improved if the authors test the impact of COVID19 pills and vaccines on their phenotypes in efferocytosis.

  7. Version published to 10.1101/2021.02.18.21251504v2 on medRxiv
  8. ScreenIT

    SciScore for 10.1101/2021.02.18.21251504: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    Institutional Review Board StatementIRB: Study approval: The procedures followed in the study were approved by the Research Ethics Committee of Hospital das Clínicas de Ribeirão Preto (CEP-FMRP/USP) and by the National Ethics Committee,
    Consent: Written informed consent was obtained from recruited patients
    Randomizationnot detected.
    Blindingnot detected.
    Power Analysisnot detected.
    Sex as a biological variablenot detected.
    Cell Line AuthenticationAuthentication: Differential expression analysis was conducted using FindMarkers function in Seurat using Wilcoxon test to compare mild and severe COVID-19 patients with healthy individuals for each cluster previously identified by authors.

    Table 2: Resources

    Antibodies
    SentencesResources
    After that, the cells were collected for evaluation of CD206 surface expression by flow cytometry (using anti-CD206 antibody clone 19.2, FITC - BD Biosciences) or gene expression analysis by RT qPCR, and the supernatant of the culture was collected for cytokine quantification (as described below).
    anti-CD206
    suggested: None
    Cells were permeabilized with 0.1% Triton-X-100 in PBS for 10 min at RT and slides were blocked with 1% bovine serum albumin (BSA), 5 µg/uL rat IgG in PBS for 45 min RT prior to overnight incubation with monoclonal rabbit anti-SARS-CoV-2 spike antibody in blocking buffer.
    anti-SARS-CoV-2
    suggested: None
    For phenotyping of blood circulating monocytes, blood samples from healthy donors and COVID-19 patients were processed for lysis of red blood cells, followed by labelling with monoclonal antibodies for CD14 (clone M5E2; BD PharmingenTM), CD16 (clone ebioCB16(CB16); eBioscience Inc., San Diego; CA, USA),
    CD14
    suggested: (BioLegend Cat# 348805, RRID:AB_2889063)
    CD16
    suggested: None
    Lung sections were labeled with primary antibodies rabbit anti-Human CD36 (ThermoFisher Scientific)
    anti-Human CD36
    suggested: None
    Then, sections were washed thoroughly in PBS and incubated for 45 min at RT with the secondary antibodies donkey anti-rabbit IgG conjugated to Alexa Fluor 488 (Jackson ImmunoResearch) and goat anti-mouse IgG conjugated to Alexa Fluor 594 (Jackson ImmunoResearch) diluted in PBS.
    anti-rabbit IgG
    suggested: None
    anti-mouse IgG
    suggested: None
    Experimental Models: Cell Lines
    SentencesResources
    Cells: Jurkat, Calu-3, Vero CCL81, and THP-1 cells were from (ATCC®).
    Jurkat
    suggested: None
    The viral stock was produced in a monolayer of Vero CCL81 cells cultured in serum-free DMEM (GIBCO) at 37°C in a 5% CO2 atmosphere.
    Vero CCL81
    suggested: None
    To generate infected apoptotic Vero CCL81 and Calu-3 cells (CoV2-AC), cells were incubated with SARS-CoV-2 at a multiplicity of infection (MOI) of 1 in serum-free media for 1h for viral adsorption, topped with fresh DMEM and incubated for 48h.
    Calu-3
    suggested: None
    For the two-round efferocytosis assay, THP-1 cells were labelled with 5 µM CellTrace™ Violet dye (Thermo Scientific).
    THP-1
    suggested: CLS Cat# 300356/p804_THP-1, RRID:CVCL_0006)
    Briefly, Violet-THP-1 cells were incubated with UV-AC-pHrodo or CoV2-AC-pHrodo for 18h at 1:1 ratio.
    Violet-THP-1
    suggested: None
    Experimental Models: Organisms/Strains
    SentencesResources
    Thymocytes were obtained from from C57BL/6 mice.
    C57BL/6
    suggested: None
    Software and Algorithms
    SentencesResources
    Images obtained by confocal and epifluorescence microscopy were analyzed using ImageJ (NIH).
    ImageJ
    suggested: (ImageJ, RRID:SCR_003070)
    Data were plotted and analyzed with GraphPad Prism 8.4.2 software (GraphPad Prism Software Inc., San Diego, CA).
    GraphPad Prism
    suggested: (GraphPad Prism, RRID:SCR_002798)

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: Please consider improving the rainbow (“jet”) colormap(s) used on pages 40 and 36. At least one figure is not accessible to readers with colorblindness and/or is not true to the data, i.e. not perceptually uniform.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  9. Version published to 10.1101/2021.02.18.21251504v1 on medRxiv