1. The dysregulation of HDL in ARDS patients.
To determine the changes of HDL in septic-ARDS, the plasma samples were collected from 40 ARDS patients and 40 matched healthy controls (Table 1 for the clinical characteristics). Compared to normal controls, ARDS patients showed significantly decreased plasma levels of HDL-C and HDL-apolipoproteins (apoA-I, apoA-II and apoA-IV, apoC-III) and the increased level of apoE. These patients also exhibited the remarkable increases of inflammatory indexes (CRP and PCT), while the reduction of paraoxonase-1 (PON1) was observed in these patients (Table 2). No significant difference in HDL-C level was observed among mild, medium and severe ARDS patients (data not show).
Following HDL isolation, the amount of important protein component was measured and the ratio to apoA-I was calculated to determine HDL composition. A plasma mixture from 5 patients with similar age and clinical situations was used for the isolation procedure to improve the quality. Compared to N-HDL, A-HDL displayed a significant decrease in apoA-I. The ratios of apoE, apoC-III and SAA to apoA-I increased in A-HDL, while the fractions of MPO, PON1, apoA-II and apoA-IV were comparable between N-HDL and A-HDL (Fig. 1A). Additionally, since deregulated apoA-1 oxidation accounted for HDL dysfunction, we compared the oxidative modifications of apoA-I in N-HDL and A-HDL by LC-MS/MS (4 HDL samples per group, 1 HDL sample isolated from 5 subjects pooled plasma). However, both groups exhibited same patterns of oxidative sites (Fig. 1B) and failed to show any significant difference in the oxidation level of each sites (data not shown).
2. The plasma HDL from ARDS patients promotes CLP-induced ALI.
Owing to the alteration in HDL composition observed in septic-ARDS patients, we further investigated the functional changes of A-HDL, by administrating N-HDL or A-HDL to C57BL/6 mice through tail vein (50 mg/kg, PBS as control), immediately after moderate CLP surgery. Although HDL treatments failed to cause obvious lung histopathologic changes and inflammation on sham mice without CLP (S-Figure 1A), the administration of A-HDL, but not N-HDL, significantly promoted CLP-induced ALI indicated by severe alveolar histopathologic disruption including thickening alveolar septum, inflammatory cells infiltration, patchy hemorrhage areas (Fig. 2A and B). A-HDL treatment also caused severe lung edema indicated by the markedly increased ratio of lung wet/dry weight (Fig. 2C). The Evens Blue leakage assay further indicated significantly aggravated pulmonary endothelial permeability by A-HDL treatment 4 hours after CLP (Fig. 2D).
The severe ALI in A-HDL treated mice was coupled with an exaggerated inflammatory response determined by the increased levels of TNF-α in BALF and the marked upregulation of TNF-α, IL-1β and MCP1 in the lung (Fig. 2E and F). Intriguingly, no difference was observed in the plasma level of LPS between mice treated by A-HDL and N-HDL, suggesting that the enhanced ALI by A-HDL was not due to abnormal increase in plasma LPS (Fig. 2G).
Given the potential effects of endogenous mouse HDL in these in vivo studies, the HDLs were administrated into apoA-I KO mice which showed massive depleted plasma HDL (Fig. 3A). These KO mice displayed severe ALI induced by moderate CLP with exacerbated pulmonary inflammation (S-Figure 2). Since moderate CLP caused high mortality on KO mice, these mice were subjected into light CLP procedure followed by HDL administration. In line with observations in C57BL/6 mice, A-HDL treatment enhanced ALI/ARDS phenotypes in apoA-I KO mice after CLP including alveolar histopathologic changes, lung permeability, lung edema and alveolar inflammation (Fig. 3B-F), while A-HDL and N-HDL treated mice showed the comparable levels of plasma LPS in these mice (Fig. 3G). These results further clearly confirmed that the adverse remodeling of HDL facilitates sepsis-induced ALI/ARDS and these deleterious effects are not due to the abnormal capability of LPS neutralization.
3. A-HDL remodeling promotes CLP-induced dysfunction of pulmonary endothelium.
To determine whether deleterious effects of A-HDL could be associated with pulmonary endothelial deregulation, we examined the adhesion proteins involved in endothelial cell-cell junction and leukocyte recruitment. CLP surgery caused significant increases in VCAM1 and ICAM1 and decrease in VE-cadherin in the lungs, whereas A-HDL treatment caused exacerbated changes suggesting a worse deregulation of pulmonary vascular endothelium (Fig. 4A and B). These findings were consistent with severe ALI/ARDS phenotype observed in these mice, suggesting that the adverse remodeling in HDL is associated with the dysfunction of pulmonary endothelium during the development of ARDS.
4. HDL from ARDS patients promotes the dysfunction of primary cultured pulmonary microvascular endothelial cells.
Since A-HDL and N-HDL treated mice show similar plasma LPS level, we reasoned that the A-HDL might have direct deleterious effects on lung vascular endothelial cells to render the lung more susceptible to sepsis-induced endothelial dysfunction. To examine this hypothesis, isolated MLECs (CD31-positive, S-figure 1B) were exposed to medium containing N-HDL, A-HDL or PBS with human albumins as control. The cells treated with A-HDL showed marked reduction of VE-cadherin and induction of VCAM1, although A-HDL treatment failed to increase ICAM1 expression (Fig. 5A). A-HDL exposure also caused marked increased expression of pro-inflammatory cytokines including TNF-α and IL-6 (Fig. 5B). Since the activation of NF-κB signaling is associated with the induction of inflammatory cytokines, the change in phospho-p65/p65 NF-kB was then measured. A-HDL treated cells exhibited a high ratio of phospho-p65/p65, whereas N-HDL exposure failed to cause activation of p65. This observation suggested that A-HDL could cause activation of pro-inflammatory signaling (Fig. 5A).
These results suggested that the dysfunction of HDL predispose the lung to sepsis-induced ALI/ARDS likely through the direct deleterious effects on endothelial cells.