Abstract
Membrane vesicles released by Pseudomonas aeruginosa PA14 can elicit septic acute lung injury due to loss of endothelial barrier integrity and inflammasome activation https://bit.ly/3gMnkPu
To the Editor:
Sepsis is a common cause of lung hyperinflammation and barrier failure, resulting in acute respiratory distress syndrome (ARDS). Despite antibiotics, mortality from bacterial sepsis increases in the developed world, suggesting injurious mechanisms beyond live bacteria. In addition to bacterial toxins, membrane vesicles (MVs) may present potential mechanisms of organ failure in sepsis. Bacterial MVs are extracellular vesicles formed from bacterial membranes [1] that can elicit systemic inflammatory responses, e.g. by inflammasome activation [2, 3]. Here, we tested whether MVs from a relevant sepsis pathogen, Pseudomonas aeruginosa, were sufficient to cause characteristic signs of acute lung injury (ALI), the preclinical analogue to ARDS, in vitro and in vivo.
MVs were collected from different bacterial strains. Equal colony forming units (CFUs) of Pseudomonas aeruginosa (PA14), Escherichia coli (Nissle 1917), Staphylococcus aureus (ATCC 6538) or Lactobacillus casei (MET-1) were inoculated into cultures for 24 h, with the exception of L. casei, which was cultured to 48 h. PA14 was additionally cultured for 6 h. Bacterial MVs were isolated by differential centrifugation and filtered [4]. MV size and number were determined by transmission electron microscopy (TEM) and flow cytometry [5]. PA14 lipopolysaccharide (LPS) content in culture (<1–3 μg·mL−1) and from isolated MVs (2 μg·mL−1) was measured (ToxinSensor Endotoxin Detection System, GenScript; Piscataway, NJ, USA). Passage 6 primary human microvascular pulmonary endothelial cells (HMPECs; C-12281 PromoCell, Heidelberg, Germany) were treated with 0–3.1×102 PA14 MVs per cell or LPS (2 μg·mL−1 of L9143; Sigma Aldrich, Oakville, ON, Canada) and trans-endothelial electrical resistance (TEER) was measured over 4 h [5]. MVs (8×1010 PA14 MVs per kg) in 200 μL of normal saline (NS) were transfused by tail vein into 12-week-old male BALB/c mice (Charles River, Canada) primed for 2 h with intraperitoneal NS or 2 mg·kg−1 LPS and MV-induced ALI was assessed after 4 h by a composite of lung wet-to-dry weight ratio, bronchoalveolar lavage fluid (BALF) protein, and a histological score [5]. For inflammasome activation, lung protein expression of NLRP3 and cleaved caspase-1 and plasma IL-1β levels were assessed. Animal experiments were approved by St. Michael's Hospital Animal Care Committee (#590).
Cultured bacteria yielded MVs of 20–200 nm in size (figure 1a and b). PA14 MVs were selected for further analysis based on PA14's known pathogenicity and the relative abundance of MVs produced by PA14 (figure 1c). Titration of PA14 MVs from 6 or 24 h culture to HMPECs revealed a loss of endothelial barrier integrity at 3.1×102 MVs per cell, while LPS at concentrations comparable to those in 3.1×102 MVs per cell caused no change in TEER (figure 1d and e). Loss of barrier function was not due to cytotoxicity, as neither LPS nor MVs caused lactate dehydrogenase release from HMPECs (figure 1f). Next, we probed whether PA14 MVs could elicit septic ALI in mice primed with a small dose of LPS. Mice exposed to LPS then MVs showed increased lung wet-to-dry weight ratio, BALF protein content and histological characteristics of ALI, while LPS or MVs alone had little or no effect (figure 1g–j). Plasma IL-1β levels, as well as lung protein expression of NLRP3 and cleaved caspase-1, were elevated in LPS+MV treated mice, indicating inflammasome activation (figure 1k–m).
Sepsis affects an estimated 49 million people and causes 11 million deaths annually [6]. Despite antibiotics, sepsis may advance to multi-organ failure and ARDS even after blood cultures become negative, suggesting excessive inflammation and/or circulating bacterial products as drivers of disease progression. Based on the emerging recognition of MVs as parts of the bacterial secretome and as carriers of virulence [7], we considered that MVs may play a direct role in the induction of ARDS/ALI in the absence of live bacteria. In line with previous studies, MVs of different bacterial species were on average ∼50 nm in size [1, 8, 9]. MV release was not related to lysis of bacteria grown to the stationary phase as 1) supernatant analyses showed no release of bacterial RNA polymerase, 2) TEM revealed no indication of bacterial lysis or debris, and 3) TEER data could be replicated with MVs from the log phase of bacterial growth (6 h in figure 1e). PA14 produced considerably more MVs than other less-pathogenic bacteria grown under similar conditions. Considering the abundance of MVs formed by PA14 and the documented pathogenicity of this strain, we examined the potential contribution of PA14 MVs to ALI. Previously, studies have implicated MVs as inducers of immune cell activation and lung inflammation [2, 10] that may contribute to the induction of ALI. Similarly, bacterial MVs have been reported to reduce monolayer integrity in extrapulmonary epithelia [11]. Here, we show that PA14 MVs cause a dose-dependent loss of endothelial barrier integrity in vitro, and induce characteristics of ALI and inflammasome activation in LPS-primed mice in vivo. Although our analyses focused specifically on ALI, one may speculate that similar mechanisms may promote extra-pulmonary organ failure in sepsis. Consequentially, means to reduce or prevent MV release e.g. by peptidylarginine deaminase inhibitors, membrane peptidogylcans or cannabidiol [8, 9, 12], or to attenuate systemic inflammation in response to MVs [13], may prove beneficial for treating septic patients. The recognition that certain MVs can act as potential inducers of microvascular barrier failure and ALI should further be considered as potential caveat in current attempts to exploit MVs as vaccine candidates against pulmonary infections [7]. This concern relates especially to patients with pre-existing inflammatory conditions mimicking the scenario of our two-hit model, and to specific bacterial species such as PA14, while MVs from other species seem to be harmless, as indicated by the fact that approved MV-related vaccines are not associated with ARDS [14].
The present study has limitations. First, while the concentrations of MVs delivered in our model are in keeping with levels of circulating extracellular vesicles in vivo, their relevance for clinical sepsis remains to be shown. Second, we only examined the effects of PA14 MVs, yet in polymicrobial sepsis various MVs may have additive or even synergistic interactions which may change ALI outcomes. Additionally, our MVs were derived from cultured cells as opposed to generated within human hosts. Future studies are required to address the implications of where and how MVs are formed. Lastly, in our in vivo model MVs only caused modest ALI in mice primed by prior LPS. Such LPS priming generates a pro-inflammatory milieu on which ALI develops once a second hit, such as transfused blood products, mechanical ventilation or bacterial infection occurs, reflecting the multifactorial pathogenesis of ARDS. Yet in vitro, even MVs alone sufficed to induce endothelial barrier failure. Notably, MVs may themselves shuttle LPS and activate toll-like receptors (TLRs) on target cells, triggering immune responses such as inflammasome activation [1, 3]. In addition to such TLR-dependent effects [2, 15], MVs may also elicit TLR-independent responses [15], indicating that mechanisms other than LPS contribute to MV pathogenicity. Consistently, in the present study TEER effects of MVs could not be replicated by equal doses of LPS alone. Exact mechanisms of MV pathogenicity in addition to MVs acting as LPS shuttles/chaperones remain to be elucidated but may comprise 1) proteolysis as described for MVs of Serratia grimesii, 2) epigenetic regulation by transfer of small RNAs, 3) MVs containing the barrier-disruptive biolipid ceramide, and 4) inflammasome activation by MVs.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01500-2021.Shareable
Acknowledgements
M.J. McVey is the recipient of a CIHR Vanier Scholarship.
Footnotes
Conflict of interest: M. Maishan has nothing to disclose.
Conflict of interest: A. Foley has nothing to disclose.
Conflict of interest: R. Turki has nothing to disclose.
Conflict of interest: E.J. Roach has nothing to disclose.
Conflict of interest: R. Deschler has nothing to disclose.
Conflict of interest: S. Weidenfeld has nothing to disclose.
Conflict of interest: N.M. Goldenberg has nothing to disclose.
Conflict of interest: C.M. Khursigara has nothing to disclose.
Conflict of interest: W.M. Kuebler has nothing to disclose.
Conflict of interest: M.J. McVey has nothing to disclose.
Support statement: W.M. Kuebler is supported by grants from the German Research Foundation (SFB-TR84 A2 and C9, SFB 1449 B1, SFB 1470 A4, KU1218/9-1, KU1218/11-1, and KU1218/12-1) and the German Ministry of Education and Research (BMBF) in the framework of SYMPATH (01ZX1906A) and PROVID (01KI20160A). C.M. Khursigara is supported by operating grants from the Canadian Institute of Health Research (CIHR; PJT 156111). N.M. Goldenberg is supported by CIHR and AstraZeneca Canada and Canadian Lung Association Early Career Scientist Awards as well as an International Anesthesia Research Society Mentored Research Award. M.J. McVey received financial support from Hospital for Sick Children Research Institute – Peter Gilgan Centre for Research and Learning start-up funds, a Canadian Hematology Society RK Smiley Research Grant and a Society of Pediatric Anesthesia Young Investigator Grant. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received May 26, 2021.
- Accepted February 4, 2022.
- Copyright ©The authors 2022. For reproduction rights and permissions contact permissions{at}ersnet.org