Fig 2: Immunohistochemical staining of the xenograft tumor sections (400X). Dark brown color was observed as positive staining. Tissues from xenograft tumors were negative for ER (A), PR (B) and HER-2 (C), and were positive for P63 (D), CK5/6 (E), and a-SMA (F) localized in the cell membrane.

Fig 3: (A) Probability of each snRNA-seq cluster being localized in the hepatic lobular region based on the DSP WTA data. (B) Representative images of serial sections from four consecutive liver core biopsies samples (BIDMC cohort, donors L1 to L4 as indicated on each column) stained for the ductular/cholangiocyte cell marker CK19, HSC activation maker a-SMA, and connective tissue (picrosirius red), as indicated. All images were acquired at the same magnification (scale bar is 50um). (C) Pathway activity score for the Kupffer cell proliferation signature described by37 for the macrophage clusters of the immune compartment. MAC4 was characterized as Replicating Kupffer Cells. (D) Pathway activity score for the endothelial cell proliferation signature described by46 for the Endothelial cell clusters. EC12 was annotated as Replicating Endothelial Cells.

Fig 4: (A) Uniform manifold approximation and projection (UMAP) for the (A) Immune / blood, (B) Endothelial cell, and (C) Mesenchymal cell compartments ((A) Immune: MAC1 n = 2,798, MAC2 n = 2,601, TC1 n = 1,522, TC2 n = 388, TC3 n = 327, TC4 n = 29, DBL1 n = 1,331, MAC3 n = 1,038, NK n = 857, PC1 n = 397, PC2 n = 124, BC n = 124, ERY-P n = 359 MAC4 n = 222, MAST n = 36 DBL2 n = 193; (B) Endothelial: EC1 n = 2,338, EC2 n = 2,247, EC3 n = 1,563, EC4 n = 1,117, EC5, n = 795, EC6 n = 379, EC7 n = 328, EC8 n = 166, EC9 n = 116, DBL3 n = 91, EC11 n = 73, EC12 n = 65, (C) Mesenchymal: MES1 n=1,223, MES2 n=1,065, MES3 n=1,040, MES4 n=374, MES5 n=328, MES6 n=312, MES7 n=275, MES8 n=30). Heatmaps capturing the expression of marker genes across the 3 distinct major compartments are displayed. (D) Heatmap portraying the cell-cell communications between the cell populations. The color gradient indicates the strength of interaction between any two cell groups. Recipient/Donor cell-type color is portrayed in a blue (healthy) to red (COVID-19) gradient, relevant to the cell composition fold-change differences between healthy and COVID-19 liver samples.

Fig 5: rVAR2-based CTC isolation using magnetic beads. a Representative confocal microscopy image of a PC3 cell captured by rVAR2-coated magnetic beads in combination with the IsoFlux™ device after spiking into 5 mL of healthy donor blood. Isolated cells were incubated with anti-cytokeratin FITC antibody (green), anti-CD45 PE antibody (red), and DAPI (blue). Scale bars, 10 µm. b A low-passage pancreatic ductal adenocarcinoma (PDAC) cell from a patient-derived xenograft captured by rVAR2-coated magnetic beads in combination with the IsoFlux™ device after spiking into 5 mL of healthy donor blood. Isolated cells were incubated with anti-cytokeratin FITC antibody (green), anti-CD45 PE antibody (red), and DAPI (blue). Scale bars, 10 µm. c CTC isolation efficiency by spike-in experiments; 100 PC3 cells or PDAC cells were spiked into 5 mL of blood and isolated by rVAR2-coated beads in combination with the IsoFlux™. The recovery was estimated by immunofluorescence staining of CK+, CD45-, DAPI+ cells (mean ± s. d., n = 3). d Recovery of PDAC cells spiked into 5 mL of blood. The number of spiked PDAC cells (10, 20, 50, 100, or 200 cells) is plotted versus the number of PDAC cells isolated by rVAR2-coated beads (n = 5). Equation shows a linear regression model