Abstract Introduction Interleukin-18 (IL-18), a marker of inflammasome activation and pyroptosis, defines a severe ARDS endotype associated with heterogeneity of treatment effects in clinical trials. Inflammasome activation and pyroptosis also release cell-free mitochondrial DNA (mtDNA), which propagates endothelial dysfunction through Toll-like Receptor 9 (TLR9). We hypothesized that sepsis plasma with high IL-18 concentrations would induce more severe endothelial injury when incubated in an established 3D human lung microvascular endothelial microphysiological system (MPS). We further hypothesized that combination therapy targeting mtDNA degradation and sensing, using DNase I and a TLR9 inhibitor (TLR9i) respectively, would attenuate endothelial inflammation, activation, and permeability, with treatment effects modified by plasma IL-18 levels. Methods We collected plasma from 47 sepsis patients enrolled in the UW Sepsis Biobank at ICU admission. IL-18 was quantified using multiplexed bead-based ELISA (Magpix), defining above-median levels as high. Each plasma sample was incubated in the MPS, untreated and separately treated with TLR9i + DNase, for 16 hours. We measured endothelial permeability (dextran diffusion) and biomarkers of inflammation (IL-1B, IL-6, IL-8, TNF-alpha, IL-18, TNF-R1), coagulation (PAI-1), and endothelial activation (ICAM-1, VCAM-1, VEGF-R1) using Magpix. We quantified mtDNA using qPCR for the mitochondrial NADH gene and DNase concentration/activity using ELISA. We compared untreated vs treated log-transformed biomarkers, stratified by IL-18-high and low groups, using paired 2-sided Student’s t tests. Results IL-18-high plasma significantly potentiated inflammatory markers (TNF-alpha 1.64 vs 1.99 log ug/uL, p = 0.05; IL-18 2.45 vs 2.67 log ug/ul, p = 0.041) compared to IL-18-low plasma, validating the endotype distinction in vitro. While treatment had minimal effect on the overall cohort, mtDNA-targeted therapy demonstrated a paradoxical, differential effect. Specifically, DNase+TLR9i improved key inflammatory biomarkers (IL-1beta, IL-6, IL-8, TNF-alpha) in IL-18-low plasma, and unexpectedly worsened permeability in IL-18-high plasma (Table 1). mtDNA concentration, DNase activity and concentration were similar between IL-18 groups. Conclusion Our MPS recapitulated endothelial dysfunction associated with high plasma IL-18 in sepsis and ARDS, suggesting utility as a physiologically relevant platform for testing endothelial-targeted therapies. mtDNA-targeted treatment paradoxically improved inflammatory biomarkers in IL-18 low but not high plasma despite comparable mtDNA and DNase concentrations. Along with clinical evidence linking elevated IL-18 to worse ARDS outcomes and heterogeneity of treatment effect, these findings suggest that Il-18 identifies a sepsis-associated endothelial endotype with variable therapeutic responses. Although further study is needed to define the underlying mechanisms, mtDNA-directed interventions may represent a beneficial therapy for the IL-18 low endotype, which currently lacks targeted therapies for sepsis and ARDS. This abstract is funded by: K08GM160929-01, KL2TR002374, and ATS CSL Behring Research Award in Acute Respiratory Distress Syndrome
Faust et al. (Fri,) studied this question.