Peptidylarginine deiminase 4 and ADAMTS13 activity in Staphylococcus aureus bacteraemia

Abstract

Staphylococcus aureus infection is associated with increased levels of neutrophil extracellular traps (NETs) and von Willebrand factor (VWF), and with reduced activity of ADAMTS13 (a disintegrin and metalloproteinase with thrombospondin type 1 motifs, member 13). Peptidylarginine deiminase 4 (PAD4) contributes to NET formation and inactivates ADAMTS13 in vitro. The role of PADs in the dynamics of NETs, VWF and ADAMTS13 has not yet been studied. We thus aimed to assess the longitudinal evolution of NETs, PADs, VWF and ADAMTS13 activity in S. aureus infection. Plasma samples from S. aureus bacteraemia patients were longitudinally collected and analysed for NETs, PAD4/PAD2, VWF and ADAMTS13 activity. Correlation analyses with clinical data were performed. Recombinant PAD4 and S. aureus were assessed in vitro for their potential to modulate ADAMTS13 activity. Sixty-seven patients were included. Plasma levels of NETs, VWF, PAD4 and PAD2 were increased and ADAMTS13 activity was decreased. Levels of PADs were negatively correlated with ADAMTS13 activity. NETs were positively correlated with PADs, and negatively with ADAMTS13 activity. In vitro, recombinant PAD4 but not S. aureus reduced ADAMTS13 activity in plasma. Levels of PAD4 and PAD2 correlate with reduced ADAMTS13 activity, with neutrophils as the likely source of PAD activity in S. aureus bacteraemia. This article is part of the Theo Murphy meeting issue 'The virtues and vices of protein citrullination'.

Keywords: ADAMTS13; Staphylococcus aureus; neutrophil extracellular traps; peptidylarginine deiminase 4; von Willebrand factor.

Figure 1.

Schematic overview of interactions between S. aureus, neutrophils, NETs, PAD4, VWF and ADAMTS13. Staphylococcus aureus infection triggers neutrophil activation and the release of NETs. PAD4 contributes to NET formation by citrullinating histones, and citrullinates ADAMTS13. This renders ADAMTS13 inactive, leading to an accumulation of ultra-large VWF multimers. NETs, neutrophil extracellular traps; PAD4, peptidylarginine deiminase 4; VWF, von Willebrand factor; ADAMTS13, a disintegrin and metalloproteinase with thrombospondin type 1 motifs, member 13. Created with Biorender.com.

Figure 2.

Longitudinal evolution of plasma levels of neutrophil counts, myeloperoxidase (MPO), C-reactive protein (CRP), citrullinated histone H3 (H3Cit), citrullinated histone H3 complexed with DNA (H3Cit–DNA), myeloperoxidase–DNA complexes (MPO–DNA), peptidylarginine deiminase 4 (PAD4) antigen, PAD2 antigen, von Willebrand factor (VWF) antigen, a disintegrin and metalloproteinase with thrombospondin type 1 motifs, member 13 (ADAMTS13). (a) Schematic overview of collection timepoints and number of plasma samples left for analysis at the indicated timepoints. Created with Biorender.com. (b) Graphs represent the individual marker measurements at the indicated timepoints. Bars indicate medians. Error bars indicate the interquartile range. Individual dots represent individual patient data. Reference ranges as used in the clinical laboratory are indicated in grey for neutrophil counts, CRP and ADAMTS13 activity. The dotted black line shows the median value of two to eight technical replicates of a normal human plasma pool, for reference of non-clinical measurements (MPO, H3Cit, H3Cit-DNA, MPO-DNA, PAD4, PAD2 and VWF). p-values were obtained from testing within a linear mixed-effects model using the lmerTest package (lmer function) in R, indicating overall evolutions over time without post hoc testing for comparisons between timepoints. PAD4 and PAD2 levels are presented on a log scale for better data visualization. (Online version in colour.)

Figure 3.

Correlation and subgroup analyses from the study data. (a) Elevated levels of PAD4 correlate with reduced ADAMTS13 activity at day 1 and between day 7 and 10. Elevated levels of PAD2 correlate with reduced ADAMTS13 activity at baseline, day 1 and between day 7 and 10. PAD4 and PAD2 levels are presented on a log scale for better data visualization. (b) Elevated levels of PAD4 and PAD2 correlate positively with neutrophil activation markers and NET biomarkers. ADAMTS13 activity shows a negative correlation with NETs. Elevated levels of VWF are associated with elevated levels of NETs. Colour gradient indicates the strength of correlations based on Spearman correlation testing. Correlations depicted in grey did not reach statistical significance. (c) Non-surviving patients have lower ADAMTS13 activity as compared to surviving patients in the late disease phase (between day 7 and 10; Mann–Whitney testing). (d) PAD4 is positively correlated with APACHE II score, a score predicting in-hospital mortality risk, early after bacteraemia onset (baseline). ADAMTS13 activity shows a negative correlation with this score at day 4 after enrolment.

Figure 4.

In vitro citrullination assay. (a) Experimental design of in vitro citrullination experiments. Four conditions were prepared to measure ADAMTS13 activity: frozen plasma, plasma incubated at 37°C for 2 h, plasma co-incubated with rhPAD4 at 37°C for 2 h, and a 50/50% mixture of the middle two conditions. (b) ADAMTS13 activity is decreased after 2 h of incubation at 37°C, and further decreases upon citrullination by rhPAD4. Bar indicates the mean percentage of two technical replicates. Error bars indicate standard deviation. (Online version in colour.)