Abstract
Background: Intravital imaging is a powerful tool for investigating dynamic biologic processes such as neutrophil (PMN) recruitment and NETs release during infection. However, the specificity of commonly used NETosis markers remains uncertain. Histone H2AX and neutrophil elastase (NE) are frequently employed in intravital microscopy, but have key limitations: both can be released independently of NETs - H2AX during apoptosis or necrosis, and NE during PMN degranulation. Additionally, both are cationic and may bind non-specifically to polyanionic structures in the endothelial glycocalyx. Citrullinated histone H3 (citH3) is a more specific marker of PAD4-dependent NETosis, but remains underutilized in intravital imaging studies.
Aim: In this study, we used intravital imaging of the hepatic microvasculature in wild-type (WT) and PAD4-deficient (Padi4-/-) mice at baseline and 4 hours post bacterial infection to define how PAD4 shapes the spatiotemporal relationship between PMNs, H2AX, NE, and citH3.
Methods: Analysis was performed on WT and Padi4-/- littermate mice on a C57Bl/6 background derived from heterozygous matings. Intravital imaging of the hepatic sinusoids, using an Olympus BX61WI microscope, was performed at baseline and 4 hours post-intraperitoneal injection of Salmonella typhimurium SL13344. Animals received i.v. injections of fluorophore-conjugated antibodies targeting F4/80 to detect hepatic resident macrophages, known as Kupffer cells (KC), Ly6G to identify PMN, and H2AX and NE to visualize putative NET components. A primary non-conjugated rabbit antibody against citH3 was injected, followed by a secondary fluorophore to visualize citH3. Quantification of histone and NE signal area, along with colocalization analysis, was performed with ImageJ software. A subset of animals received the pan-PAD inhibitor, BB-Cl Amidine (10 mg/kg/day), prior to bacterial injection. Another group was treated with an anti-Gr-antibody (200 µg/day) to deplete PMN before Salmonella exposure.
Results: Prior to Salmonella injection, few Ly6G-positive PMN were observed. However, H2AX was present in the sinusoids of both WT and Padi4-/- mice, with a similar area of distribution. NE was also observed in equal amounts in WT and Padi4-/- mice and highly colocalized with H2AX. CitH3 was also detected in both Padi4-/- and WT animals, with significantly lower staining area in Padi4-/- mice compared to WT animals (p<0.01). While H2AX and NE formed a continuous thin layer lining the sinusoids, citH3 staining was more discrete, with rounded protrusions into the sinusoids. The few PMNs colocalized with citH3, but not with H2AX or NE.
Post-Salmonella injection, recruitment of Ly6G-positive PMN to the liver increased in both WT and Padi4-/- mice. In WT mice, this was accompanied by a marked rise in both H2AX and NE (p<0.01), with a similar spatial distribution of the two markers. In Padi4-/- mice, H2AX levels also rose significantly, though they remained slightly below those observed in WT animals. Notably, there was no corresponding increase in NE in Padi4-/- mice. CitH3 levels increased in both WT and Padi4-/- mice after Salmonella injection, though levels were significantly lower in Padi4-/- mice (p<0.01). In both genotypes, citH3 did not colocalize with H2AX or NE. Notably, at the time of imaging at 4 hours after bacterial injection, most citH3 deposited in the sinusoids had been phagocytosed by KCs, demonstrating KCs' crucial role in rapidly clearing a major NET component. Intriguingly, H2AX and NE were not phagocytosed by KCs. Treatment with the pan-PAD inhibitor BB-Cl Amidine reduced citH3 and NE levels post-Salmonella in both WT and Padi4-/- mice. PMN depletion was associated with reduced, but not eliminated, sinusoidal H2AX and citH3, suggesting these proteins were released under homeostatic conditions, prior to PMN depletion.
Conclusions: Our study demonstrates that citH3 release in the hepatic microvasculature during bacterial infection is predominantly but not exclusively PAD4-dependent. In contrast, H2AX and NE are released via PAD4-independent mechanisms and persist in the vasculature despite clearance of citH3 of KCs, suggesting distinct regulatory and clearance pathways for citH3 vs. H2AX and NE. These findings reveal potential limitations of PAD4 inhibition as a therapeutic strategy to limit NET-associated microvascular injury.
