Abnormal GC architecture in LDLNs of COVID-19 decedents correlates with lymphatic clotting in LDLN. (A-D) Representative images from stained LDLNs comparing features of normal (left), regressed (middle), and otherwise abnormal (right) GC follicles of COVID-19 and control decedents. (A) H&E-stained tiled images where the abnormal (right) lacks secondary follicle formation altogether. (B) Zoomed-in H&E-stained images showing a normal secondary follicle (left), a regressed follicle lacking a mantle zone and containing TBMs (black arrows) (middle), and complete lack of follicle architecture (right). (C) Representative immunofluorescence images showing lymphocyte activation in control (left) and COVID-19 (middle and right) LDLNs; CD83 (cyan), GL7 (red), and CD20 (green). The left panel shows strong follicle formation and activation, the middle panel shows decreased follicle size and density, and the right panel shows abnormal follicle structure and extensive extrafollicular activation. (D) Representative immunofluorescence images of T-cell (CD3, red) and B-cell (CD20, green) zone integrity in COVID-19 LDLNs. The left panel shows dense B cells and distinct T-cell and B-cell separation, whereas the middle panel shows decreased follicle size and density, and the right panel shows decreased cellularity and poor T-cell and B-cell zone integrity. (E) Average GC abnormality scores for LDLNs of COVID-19, H1N1, and control decedents. Dashed line represents the median for each group, and dotted lines represent the first and third quartiles. ∗∗∗P < .001, one-way ANOVA test with Tukey multiple comparisons test. (F) Linear regression correlations of GC abnormality score vs percentage of clotted lymphatic vessels in LDLNs from COVID-19 (black), H1N1 (blue), and control (red) decedents. Scale bars, 500 μm (A) and 100 μm (B-D).