Fig 2: Pathological features of autopsy myocardial tissue from SARS-CoV-2-infected patients.A. Hematoxylin and Eosin (H&E) staining of healthy neonatal left ventricle tissue. B. H&E staining images of myocardial tissue from a COVID-19 patient with diagnosed myocarditis (COVID-M1). Black boxes indicate zoomed in areas of the bottom panels. Red arrowheads indicate cardiomyocytes lacking chromatin stain. C. Representative H&E staining images of myocardial tissue from a COVID-19 patient without diagnosed myocarditis (COVID-A1). Red arrowheads denote putative nuclear locations lacking chromatin stain. D Representative immunofluorescence staining of the myocardial tissue from a COVID-19 myocarditis patient (COVID-M1). Cardiomyocytes show diffuse and disorganized cardiac troponin T (cTnT) staining with occasional cells in the blood vessel staining positively for cTnT (green). White boxes indicate zoomed in areas of the bottom figures. White arrowhead indicates cardiac troponin T material in the cytoplasm of a mononuclear cell within a blood vessel. E. Regions of COVID-A1 patient heart depicting normal (left) and disrupted (right) sarcomere structure. White boxes indicate zoomed in areas of the figures on the right. White arrows indicate breaks in a-actinin 2 (ACTN2) staining.

Fig 3: Immunofluorescence characterization of cardiomyocytes differentiated from the cardiac progenitor cells. A–D. The cardiomyocytes formed well-defined cross-striated sarcomeric structures as determined by the expression and spatial organization of Actn2, Tnnt2, and Myh6 and also expressed the ventricular specific protein Myl2. E–F. The cells also formed robust intercellular electrical and mechanical connections as determined by the spatial organization of Cdh2 and Gja1. G. RFP(+) cardiomyocytes stained positive for sodium/potassium ATPase (Atp2a2).

Fig 4: ZEB2 is expressed in injured cardiomyocytes.a Study design. b Area of the heart subjected to single-cell sequencing. c-d t-SNE map indicating transcriptome similarities among all individual cells. c Different numbers and colors highlight different clusters identified by K-medoids clustering. d Colors highlight the conditions of the hearts from which the cells were derived (sham in green and 3 dpMI in pink). e The PANTHER Classification System analysis on enriched genes from injured cardiomyocytes when compared to sham. f t-SNE map showing the expression of Zeb2 across all cardiomyocytes from sham and 3 dpMI hearts. g Representative confocal images from sham and ischemic mouse hearts 3 dpMI stained for ZEB2, ACTN2, and DAPI. h Representative confocal images from human control and ischemic hearts stained for ZEB2, ACTN2, and DAPI. Expression is depicted as normalized transcript count on a color-coded scale. t-SNE indicates t-distributed stochastic neighbor embedding. White arrows show ZEB2 positive cardiomyocytes.

Fig 5: Advantages of using small probes in imaging. (A) 2D STED image of a late cytokinetic furrow in cells labelled using an antibody to acetylated tubulin (T7451, Sigma-Aldrich, shown in magenta; signal is from a combination of primary and secondary antibodies) and the tubulin Affimer (Affimer 32, shown in green). These images, summarised in the diagram below, show how densely packed microtubules (25 nm wide) in the central region of the midbody (so-called ‘cut-zone’, arrowed in the STED images) are labelled by the smaller Affimer probe (less than 4 nm long), but not by the more commonly used combination of primary and secondary antibodies (each ∼15 nm long) (see also Tiede et al., 2017). (B) 2D STED image of Z-discs in cardiac muscle tissue, labelled with an antibody to the Z-disc protein ACTN2 (A7732, Sigma-Aldrich; ACTN2) using a combination of primary and secondary antibodies (magenta), and an Affimer to ACTN2 (green). The diagram below illustrates the structures of F-actin, full-length ACTN2 (PDB 4D1E), the ACTN2 Affimer bound to the calponin homology (CH) domains of ACTN2 (PDB 6SWT) (Curd et al., 2021) and primary and secondary antibodies (PDB 1IGY) drawn to scale, for part of the Z-disc. The anti-parallel ACTN2 dimer crosslinks actin filaments (derived from sarcomeres on either side). The combination of primary and secondary antibodies (>20 nm in length) is less able to penetrate the full width of the Z-disc (see arrows on STED images) as illustrated here, accounting for reduced ACTN2 staining in the central region of the Z-disc. The Affimer is much smaller and able to penetrate the entire Z-disc. The restricted localisation of antibodies is noticeable in STED, with a resolution of ∼50 nm, but would not be seen in confocal microscopy, with a resolution of ∼250 nm.