Endothelialitis, Microischemia, and Intussusceptive Angiogenesis in COVID-19

Abstract

COVID-19 has been associated with a range of illness severity-from minimal symptoms to life-threatening multisystem organ failure. The severe forms of COVID-19 appear to be associated with an angiocentric or vascular phase of the disease. In studying autopsy patients succumbing to COVID-19, we found alveolar capillary microthrombi were 9 times more common in COVID-19 than in comparable patients with influenza. Corrosion casting of the COVID-19 microcirculation has revealed microvascular distortion, enhanced bronchial circulation, and striking increases in intussusceptive angiogenesis. In patients with severe COVID-19, endothelial cells commonly demonstrate significant ultrastructural injury. High-resolution imaging suggests that microcirculation perturbations are linked to ischemic changes in microanatomic compartments of the lung (secondary lobules). NanoString profiling of these regions has confirmed a transcriptional signature compatible with microischemia. We conclude that irreversible tissue ischemia provides an explanation for the cystic and fibrotic changes associated with long-haul COVID-19 symptoms.

Figure 1.

Data from a community-based cohort study identified age- and sex-adjusted relative risk (RR) for hospitalizations (A) (Batty and Hamer 2020). The RR for hospitalizations in this cohort correlated closely with the Framingham risk score (quintiles) (R² = 0.997) (B) (D'Agostino et al. 2008). The error bars reflect 95% confidence intervals.

Figure 2.

Chest computerized tomography (CT) imaging in patients with documented COVID-19 disease. (A–D) Frontal images demonstrating patchy inflammatory infiltrates in the peripheral lung. (E–F) Transaxial images showing a similar appearance with inflammatory opacities revealing secondary lobules.

Figure 3.

Autopsy anatomy of patients succumbing to COVID-19. (A) Autopsy image of a COVID-19 lung with cystic changes (black arrow) and intravascular clot (white arrow) grossly apparent (scale bar, 1 cm). (B) Microscopic images of H&E-stained lung demonstrated dilated alveolar ducts (scale bar, 200 μm). (C) Wet weight of the lung in autopsy lungs from COVID-19, influenza, and uninfected control lungs (mean ± 1 SD) (Ackermann et al. 2020e). H&E histopathology of the peripheral lung demonstrating vascular blood clots in small vessels (D, white arrow; scale bar, 200 μm) and alveolar capillaries (E, black arrows; scale bar, 50 μm). (F) The relative frequency of thrombi in pulmonary arteries, alveolar capillaries, and veins (mean ± 1 SD) (Ackermann et al. 2020e).

Figure 4.

Endothelialitis and endothelial injury. Corrosion casts of control (A) and COVID-19 (B) lungs (scale bars, 100 μm) examined by scanning electron microscopy (SEM). (C) SEM of intussusceptive pillars in COVID-19 (scale bar, 10 μm). (D) Transmission electron microscopy demonstrating virus-like particles in the endothelial cells (yellow ellipse; scale bar, 10 μm). (E) Immunohistochemistry with antispike protein monoclonal antibody staining of COVID-19 endothelial cells (arrow). (F) SARS-CoV-2 probes demonstrating the presence of viral nucleic acids by fluorescent in situ hybridization (FISH) (arrows) (Ackermann et al. 2020b).

Figure 5.

Pulmonary immune response. (A–C) Immunohistochemistry of the perivascular mononuclear infiltration in the perivascular, interstitial, and alveolar compartments of the COVID-19 lung. Cells were characterized as T-cell (CD3) subsets (CD4, CD8), myeloid cells (CD15, CD68), and B cells (CD20). (D) Multiplex fluorescence immunohistochemistry identified the inflammatory infiltrate to consist predominantly of T cells: CD3 (white), CD4 (green), CD8 (yellow), CD20 (purple), CD68 (red), and DAPI nuclear stain (blue). Clot is noted in the vessel lumen (arrow; scale bar, 100 µm). (E) A CD4⁺ lymphocyte migrates into interstitial lung lymphatics (arrow; scale bar, 10 µm).

Figure 6.

Summary of NanoString transcriptional profiles of patients with COVID-19 and influenza. (A) Randomly colored projection of secondary lobular anatomy superimposed on a whole-lung synchrotron section. (B) Synchrotron images of a core lung biopsy of a COVID-19 autopsy lung. The portion of the biopsy demonstrating fibrotic remodeling (a) is colored yellow for presentation purposes (image courtesy of Dr. Claire Walsh and colleagues at the Centre for Advanced Biomedical Imaging, University College, London, UK). Scale bar, 2 mm. The secondary lobule with preserved anatomy is juxtaposed (b). (C) Higher resolution image shows the electron-dense, blood-filled vessel at significantly higher density than in the remodeled region (inset). (D) A NanoString Venn diagram reflecting the amount of shared gene expression between COVID-19 autopsy lungs and autopsy lungs obtained from patients with usual interstitial pneumonia (UIP), nonspecific interstitial pneumonia (NSIP), and alveolar fibroelastosis (AFE). (E) The NanoString profiles comparing patients with COVID-19 and influenza (column 1 and 2). Also shown are the profiles of COVID-19 changes over their hospital course (column 3) and the transcriptional profiles of COVID-19 samples taken from patients with high levels of intussusceptive angiogenesis (IA). The transcriptional profiles are summarized by the false discovery rate (FDR) and normalized pathway enrichment score (NES). (F) Schematic summary of the bronchial circulation demonstrating the intercommunication with the pulmonary circulation at the level of the secondary lobule.