Leukocytic ADAM10 and ADAM17 modulate disease severity and systemic outcome in bacterial and viral pneumonia

Graphical abstract

A disintegrin and metalloproteinase (ADAM)10 and ADAM17 modulate inflammation and immune cell function during viral and bacterial pneumonia, with local effects in the lung and potential systemic consequences via exosomes. TNF: tumour necrosis factor; IL: interleukin; IL-6R: IL-6 receptor; ROS: reactive oxygen species; COVID-19: coronavirus disease 2019.

Abstract

Background

Pneumonia caused by viral or bacterial pathogens such as severe acute respiratory syndrome coronavirus 2 or Pseudomonas aeruginosa may result in life-threatening disease with a strong contribution of protease dysregulation. The present study aimed to systematically characterise the contribution of ADAM10 and ADAM17 on leukocytes and circulating exosomes to viral and bacterial pneumonia.

Methods

The analysis of coronavirus disease 2019 (COVID-19) and bacterial pneumonia patient samples was combined with in vivo experiments in conditional knockout animals lacking either ADAM10 or ADAM17 in leukocytes and cell culture experiments for mechanistic studies.

Results

Hospitalised bacterial pneumonia and COVID-19 patients displayed a severity dependent increase of ADAM10 and ADAM17 activity on exosomes. These exosomes caused pathophysiological changes of cardiomyocytes and the endothelial barrier. In a pre-clinical murine pneumonia model, we observed that leukocytes contributed to this increase in exosomal proteolytic activity. In the local environment of the lung, ADAM10 orchestrated a pro-inflammatory response with M1 macrophage polarisation, increased reactive oxygen species (ROS) generation, cytokine release, tissue damage and oedema formation, whereas ADAM17 seemed to dampen the initial inflammatory response to an anti-infective, ROS-balanced level.

Conclusion

Leukocytic ADAM10 and ADAM17 and their release on exosomes may constitute relevant regulatory elements in bacterial and viral pneumonia, with a potential contribution of exosomes to disease progression and systemic inflammatory responses. Therefore, the diagnostic, prognostic and therapeutic value of ADAM10 and ADAM17 should be evaluated in further preclinical and translational studies, addressing the changes of the immune response and exosomes as cargo vehicles both at local site and for the prevention of systemic effects.

Shareable abstract

Leukocytic ADAM10 and ADAM17 and their release on exosomes modulate the local and systemic response upon bacterial and viral pneumonia and should be evaluated for diagnostic, prognostic and therapeutic options in future translational and clinical studies https://bit.ly/4pQSGXI

Introduction

Pneumonia caused by bacterial or viral pathogens such as Pseudomonas aeruginosa or as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1, 2] may result in a life-threatening respiratory disease striking mainly the alveoli and lower bronchioles of lungs [3]. In both cases, pulmonary infection can lead to severe innate immune cell infiltration into the bronchial airways and the alveolar space [4]. The extravasation of monocytes and neutrophils into the pulmonary tissue along with their cationic polypeptides, enzymes, as well as the production of reactive oxygen species (ROS) play a detrimental role in pathogen killing and elimination of the infection [5]. Although inflammatory cell recruitment aims to control and clear the infection, massive recruitment can end up with organ damage and substantially contribute to mortality [6].

During the past decade it has become evident that the protease web of serine-, cysteine-, metallo-, aspartic- and threonine proteases with their complex substrate spectrum, matrix-degrading activity, and interconnectivity are key components in pulmonary pathogenesis [7]. One superfamily of the metalloproteinases are metzincins, including the families of matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinases (ADAMs). ADAM proteases are transmembrane, zinc-dependent metalloproteinases, regulating the release of soluble ectodomains of a wide number of inflammatory mediators by ectodomain shedding close to the plasma membrane. Besides their catalytic activity, ADAM proteases contain a disintegrin-like domain fulfilling vital functions in cell adhesion, signalling and migration [8, 9]. In particular, ADAM10 and ADAM17 hold crucial functions during infectious diseases ranging from pathogen recognition, entrance, leukocyte recruitment and barrier destruction to resolution of inflammation (reviewed in [10]). ADAM17, also known as tumour-necrosis factor (TNF)-α-converting enzyme (TACE), mediates TNF-α, l-selectin and interleukin-6 receptor (IL-6R) shedding on leukocytes, thus critically contributing to septic shock induced by lipopolysaccharide and Escherichia coli-mediated peritoneal sepsis [11–13]. Epithelial upregulation and activation of ADAM10 and ADAM17 during P. aeruginosa infection may promote leukocyte recruitment and oedema formation through cleavage of junction molecules and regulation of cell survival and leukocyte adhesion [14, 15]. During the research on SARS-CoV-2 ADAM proteases, cleaving angiotensin-converting enzyme (ACE)2 or the spike protein [16–18] emerged as suitable drug targets to prevent a cytokine storm or tissue damage in pneumonia also related to cardiovascular side-effects (reviewed in [10]). Feasible circulating cargo vehicles mediating this long-distance communication could be either leukocytes themselves or extracellular vesicles [19]. Extracellular vesicles are a heterogenous populations of nano-to micro-sized vesicles in the range of 30–10 000 nm [20] actively triggering genetic and/or metabolic responses [21]. Exosomes have been shown to carry proteases on their surface and to be actively involved in extracellular remodelling [22], especially within the tumour microenvironment [23]. However, a comparative study of the single contribution of leukocytic ADAM10 and ADAM17 to pneumonia and potential systemic effects is missing so far.

In the present study, we found a severity-dependent increase of ADAM10 and ADAM17 activity on exosomes in bacterial pneumonia and coronavirus disease 2019 (COVID-19) patient samples, respectively. These proteolytic active exosomes caused changes of the calcium amplitude of cardiomyocytes and led to endothelial barrier disruption at least in vitro. In a pre-clinical pneumonia model, we observed that leukocytes critically contribute to this increase and that leukocytic ADAM10 and ADAM17 may drive the severity of P. aeruginosa-induced pneumonia through the regulation of macrophage polarisation, phagocytosis and bacterial clearance, ROS generation, tissue infiltration and oedema formation. Thus, leukocytic ADAM10 and ADAM17 and their release on exosomes may constitute relevant regulatory elements in bacterial and viral pneumonia and disease severity. Especially, the (mechanistic) contribution of exosomes to severe side-effects at distinct organ sides and their diagnostic, prognostic and therapeutic value should be addressed in further translational and pre-clinical studies.

Material and methods

Antibodies, inhibitors and substances

Details are provided in supplementary table S1.

Ethical approval

The use of human samples was approved by the ethics committee of the Landesärztekammer des Saarlandes (172/18 by D. Yildiz and 62/80 by R. Bals) and performed in accordance with the Declaration of Helsinki. Informed consent was obtained from all individuals. Details of patient characteristics are provided in supplementary table S2. Animal experiments were approved by the Saarland animal welfare and ethics committee (project identification number 2.4.2.2-42/2018 by D. Yildiz). For details, refer to the supplementary material.

Mice

Vav-Adam10^−/− or Vav-Adam17^−/− mice were bred as described previously [11, 24]. Details are provided in the supplementary material.

Exosome preparation

Plasma of Vav-Adam10^−/− and Vav-Adam17^−/− mice, serum of healthy, pneumonia or COVID-19 patients and the cell lysates of THP-1 cells or human neutrophils with and without infection with P. aeruginosa were subjected to extracellular vesicles and exosome purification as described in [15]. Details are provided in the supplementary material.

Electron microscopy

THP-1 cells were infected with P. aeruginosa 103 (2 h, multiplicity of infection 5), and the exocytosis of the extracellular vesicles was visualised by electron microscopy. Details are provided in the supplementary material.

Bacteria preparation

P. aeruginosa 103 or P. aeruginosa PAO1-GFP were prepared as described previously [15]. Details are provided in the supplementary material.

P. aeruginosa-induced lung pneumonia model and sample processing

Mice were anaesthetised by intraperitoneal injection of ketamine (100 mg·kg⁻¹)/xylazine (25 mg·kg⁻¹) followed by intranasal instillation of P. aeruginosa 10⁵ CFU in 20 µL PBS (10 µL per nostril) or PBS as a control and probed after 12 h, as described previously [25]. For description of CFU determination, flow cytometry analysis of bronchoalveolar lavage (BAL) cells, cytokine measurements, histochemistry and immunohistochemistry, refer to the supplementary material.

Cell culture

The isolation of human primary neutrophils and rat cardiomyocytes as well as the culture of vascular endothelial cells and the respective treatments are described in the supplementary material.

Calcium imaging and trans-endothelial electrical resistance

Calcium imaging and trans-endothelial electrical resistance (TEER) measurements are described in the supplementary material.

Western blot

THP-1 cells, human neutrophils, and exosomes were subjected to Western blot analysis as described previously [15]. For details, refer to the supplementary material.

ROS production

ROS production was quantified using Cellular ROS detection kit (Abcam, Cambridge, UK) following the manufacturer's protocol. For details, refer to the supplementary material.

Phagocytosis, cell survival and killing capacity

THP-1 or human neutrophils were pre-treated and infected with green fluorescent protein (GFP)-labelled P. aeruginosa (PA01) for 2 h and subsequently subjected to flow cytometric analysis and plating, as described in the supplementary material.

Fluorescence resonance energy transfer-based activity measurements

Screening of ADAM10 or ADAM17 activity was done using polypeptides with specific substrate cleavage sites and fluorescence resonance energy transfer linked fluorophores. Details are provided in the supplementary material.

Mass spectrometry

The supernatants and lysates of vascular endothelial cells were subjected to mass spectrometric analysis upon stimulation with exosomes or control stimulation. Details are provided in the supplementary material.

Statistics

Quantitative data are shown as mean±sd of at least three independent experiments and as indicated in the figure captions. Statistical analysis was performed using GraphPad PRISM 9.0 (GraphPad Software, La Jollla, CA, USA). A p-value <0.05 was regarded significant. Detailed analysis is described in the figure captions.

Results

Pulmonary infection causes systemic releases of ADAM proteases on exosomes correlating with disease severity and changes at distinct organ sites

Extrapulmonary complications are common in patients with pneumonia, including sepsis, increased cardiovascular risk and a decline in functional status [3]. One potential cargo vehicle of these interorgan circuits could be exosomes. Therefore, we analysed the amount of active ADAM10 and ADAM17 on exosomes (hereinafter referred to as exosomal activity) in the serum of healthy volunteers and those with bacterial pneumonia, respectively. Indeed, bacterial pneumonia led to an increase of both ADAM10 and ADAM17 activity on circulating exosomes (figure 1a,b; for kinetics and raw data refer to supplementary figure S1a,b,e,f). Next, we questioned whether this increase is specific for bacterial pneumonia patients or a general response in hospitalised pneumonia patients. We observed an increase in exosomal ADAM10 and ADAM17 activity with a severity-dependent increase in activity for ADAM17 as indicated by a significantly higher exosomal activity in COVID-19 patients with severe symptoms (requiring intensive care unit admission) compared to mild symptoms (figure 1c,d, supplementary figure S1c,d,g,h). ADAM10 and ADAM17 might be released as soluble ectodomains by the cleavage of ADAM8; however, neither ADAM10 nor ADAM17 proteolytic activity was observed in the supernatants of the extracellular vesicle and exosome preparations (supplementary figure S1i,j). The COVID-19 pandemic showed that local infections cause disease at distinct organ sites without direct infection [26], and that sole inhibition of the cytokine storm is not sufficient as curative treatment [27, 28]. One possible scenario potentially causing severe side-effects could be the release of ADAM proteases and their activity on exosomes, with vessel lining endothelial cells and cardiomyocytes as potential receiver cells. Exosomes from pneumonia patients increased the calcium amplitudes in primary cardiomyocytes and disrupted the endothelial barrier integrity (measured by TEER), effects that were prevented by inhibiting ADAM10 or ADAM17 proteolytic activity on the exosomes (figure 1e,f). Exosomes from healthy volunteers had no measurable impact on both functions. Untargeted mass spectrometric analysis of the cell supernatant of primary lung endothelial cells revealed >900 differentially shed or released proteins (figure 2a). Exosomes from pneumonia patients shed cadherins, cellular adhesion molecules and integrins (figure 2b). The differences in, for example, vascular endothelial (VE)-cadherin (ADAM17 substrate) and secreted protein acidic and cysteine-rich (SPARC) were directly reflected in the functional TEER measurements (figure 1f, figure 2b). Furthermore, exosomes from healthy volunteers led to increased release of angiogenesis promoting and blood pressure regulating proteins (e.g. angiotensinogen) as well as immune modulating decoy receptors (e.g. urokinase plasminogen activator receptor (CD87/PLAUR)) (figure 2b,c). In contrast, exosomes from pneumonia patients finally led to induction of matrix degrading enzymes (e.g. MMP14), antifibrinolytic activity (e.g. plasminogen activator inhibitor type 2 (PAI-2/SERPINB2)) and apoptosis (e.g. annexins, caspase 3); the latter eventually due to the high cleavage rate of junction molecules (figure 2c). Thus, ADAM proteolytic activity on circulating exosomes may be indicative for systemic responses during lung infection.

FIGURE 1.

a–d) The sera of healthy, pneumonia or coronavirus disease 2019 (COVID-19) patients were diluted in Hanks’ balanced salt solution (1:3). The supernatants were subjected to differential centrifugation (isolation of extracellular vesicles) and sucrose gradient centrifugation (purification of exosomes). The pellet was resuspended in DMEM medium (phenol red free) for fluorescence resonance energy transfer activity measurements with a, c) a disintegrin and metalloproteinase (ADAM)10- and b, d) ADAM17-specific probes. Data are presented as mean±sd of the net activity (n=5). e, f) Exosomes were isolated from the blood of pneumonia patients or healthy volunteers and used for the stimulation of primary rat cardiomyocytes or lung endothelial cells for 2 h. ADAM10 or ADAM17 were inhibited on the exosomes (wash step to remove free inhibitor prior to stimulation) or exosomes were left untreated as control. Cardiomyocytes were subjected to e) calcium imaging, and the effect on endothelial cells was investigated by f) trans-endothelial electrical resistance (TEER) measurement. ns: nonsignificant; ICU: intensive care unit. *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001, one-way ANOVA and Tukey post-test.

FIGURE 2.

a–c) Exosomes were isolated from the blood of pneumonia patients (n=3 for each experiment) or healthy volunteers (n=3 for each experiment) and used for the stimulation of primary lung endothelial cells for 2 h. Untargeted proteomic analyses were performed for the supernatant of cells incubated with PBS (control stimulus and solvent of exosomes), exosomes derived from the serum of healthy patients or exosomes derived from the serum of pneumonia patients. a) Venn diagram; b) heatmap of a disintegrin and metalloproteinase (ADAM)10/ADAM17 substrates; c) downstream pathways (n=2). VE: vascular endothelial; PLAUR: urokinase plasminogen activator receptor; CASP: caspase; SPARC: secreted protein acidic and cysteine-rich; MMP: matrix metalloproteinase; SERPINB2: plasminogen activator inhibitor type 2.

Leukocytes may be one essential source for the systemic release of ADAM10/17-positive exosomes upon lung infection

Epithelial cells and leukocytes build the first line of defence against lung infection. It could be argued that the release of these epithelial-derived exosomes [14, 15] to the circulation would be difficult. Therefore, we investigated the release of exosomes by leukocytes in both primary human neutrophils and THP-1 cells (monocytic cell line) and observed a cell-associated increase of ADAM10 and a decrease of ADAM17 protein expression upon infection with P. aeruginosa (supplementary figure S2a,b). The formation and exocytic release of exosomes upon infection with P. aeruginosa was associated with a strong release of mature ADAM10 in these exosomes which was absent in control-treated cells (figure 3a,b, supplementary figure S2c,d). Mature ADAM17 was constitutively released in exosomes and increased 1.5-fold upon infection for both cell types (figure 3c,d,f, supplementary figure S2c,d). In addition, only THP-1 cells showed a release of the ADAM17 pro-form on exosomes (figure 3d). P. aeruginosa stimulation led to a general, approximately two-fold increase of exosome release (figure 3a–d, supplementary figure S2e,f). The quantification of the amount of ADAM10 and ADAM17 released on exosomes in comparison to the cell-associated expression (cell lysate) revealed a constant proportion of 1% for ADAM17, whereas the proportion of ADAM10 released on exosomes was increased up to 40% upon infection (supplementary figure S2e,f, exemplary electron microscopy images supplementary figure S3a). To gain insights into the in vivo relevance of these findings, we analysed the systemic release of exosomes in a pre-clinical murine infection model (P. aeruginosa PA103 infection for 12 h). Infection of wild-type (littermate) mice with P. aeruginosa significantly increased the activity and expression (only mature form detectable) of Adam10 on exosomes, while this increase was completely blocked in mice with Adam10-deficiency in leukocytes (Vav-Adam10^−/− mice, supplementary figure S3b–e). Furthermore, the deficiency of Adam10 in leukocytes reduced the basal level of Adam10 on exosomes by 40%. In comparison to Adam10, the activity and expression of Adam17 on circulating exosomes were only slightly increased upon infection (figure 3e,f). However, mice with Adam17-deficiency in leukocytes (Vav-Adam17^−/− mice) showed a general reduction of Adam17 activity on exosomes by 50%. Thus, leukocytes may essentially contribute to the systemic release of ADAM proteases and their activity on exosomes during lung infection.

FIGURE 3.

a–d) THP-1 cells or human neutrophils were infected with Pseudomonas aeruginosa (PA103; multiplicity of infection 5) for 2 h. The supernatants were subjected to differential centrifugation (isolation of extracellular vesicles (EVs)) followed by sucrose gradient centrifugation (purification of exosomes). Exosomes were analysed for expression of a disintegrin and metalloproteinase (ADAM)10 (C-terminus), ADAM17 (C-terminus), flotillin-1 and CD9 by Western blot. Data are shown as representative images of three independent experiments. e, f) Appropriate littermates, Vav-Adam10^−/− or Vav-Adam17^−/− mice were infected with 10⁵ CFU of P. aeruginosa by intranasal instillation and probed after 12 h. Collected plasma was diluted in Hanks’ balanced salt solution (1:3). The supernatants were subjected to differential centrifugation (isolation of EVs) and sucrose gradient centrifugation (purification of exosomes). The pellet was resuspended in DMEM medium (phenol red free) for fluorescence resonance energy transfer activity measurements with a) ADAM10- and b) ADAM17-specific probes. Data are presented as mean±sd of the slope of the activity (n=5). ns: nonsignificant. ****: p<0.0001, one-way ANOVA and Tukey post-test.

Leukocytic ADAM10 and ADAM17 differentially influence disease severity and change the local immune response in P. aeruginosa-induced pneumonia

Littermate, Vav-Adam10^−/− and Vav-Adam17^−/− mice were further analysed for local and systemic inflammatory parameters. Control-treated animals did not display any basic or genotype-dependent phenotype, as shown previously [29]. Vav-Adam10^−/−mice showed a lower pneumonia manifestation, as indicated by a lower Mouse Clinical Assessment Scoring System for Sepsis (M-CASS) (supplementary table S3) score, and reduced hypothermia (figure 4a). In contrast, Vav-Adam17^−/− mice displayed an increased M-CASS score with enhanced hypothermia in comparison to infected littermates (figure 4b). Infected Vav-Adam10^−/− mice were protected against oedema formation (lung wet/dry ratio and thickness of alveolar septa) and barrier disruption as measures for disease severity, whereas infected Vav-Adam17^−/− mice showed an increase of all parameters (figure 4c–e). Leukocyte recruitment is one further hallmark of pneumonia not only enabling for proper bacterial clearance, but also causing tissue damage in case of an overspilling immune response. Flow cytometric analyses of alveolar leukocytes revealed similar numbers of neutrophils and macrophages, but less recruitment of monocytes in Vav-Adam10^−/− mice (figure 5a; for gating strategy refer to supplementary figure S4a). This was associated with a decrease of the pro-inflammatory cytokines TNF-α, CXCL1 and IL-6 as well as the soluble (s)IL-6R in the BAL compared to infected littermate mice (supplementary figure S4b). In contrast, all cell types were reduced in infected Vav-Adam17^−/− mice (figure 5b). Based on the Adam17-deficiency in leukocytes, these mice showed decreased TNF-α levels, but no differences in CXCL1, IL-6 or sIL-6R levels (supplementary figure S4c). Similar effects were observed at the systemic level. While Adam10-deficiency in leukocytes reduced the systemic increase of TNF-α and sIL-6R upon infection, TNF-α, but not sIL-6R, was decreased by Adam17-deficiency (supplementary figure S5a,b). Interestingly, both Vav-Adam10^−/− and Vav-Adam17^−/− mice displayed leukocytosis with increased blood neutrophil and monocytic cell numbers and decreased CFU numbers in the blood compared to infected littermate mice (figure 5c,d and figure 6a,b). Increased severity could be due to a higher bacterial burden in the lung. Indeed, Adam17-deficiency in leukocytes resulted in a strong increase of the BAL CFUs (figure 6c), fitting with the reduced recruitment of defending leukocytes. Surprisingly with respect to the equal neutrophil and macrophage numbers, but fitting with less disease severity, Vav-Adam10 ^−/− mice showed much higher bacterial clearance (figure 6d). This was further reflected at the tissue level, with no effect in Vav-Adam17^−/−, but reduced CFU numbers in Vav-Adam10^−/− mice (figure 6e,f). Thus, leukocytic Adam10 and Adam17 seem to be important regulators of the local immune response in the lung, resulting in protection by Adam10-defiency in leukocytes, but increasing severity upon Adam17-deficiency.

FIGURE 4.

Appropriate littermates, Vav-Adam10^−/− or Vav-Adam17^−/− mice were infected with 10⁵ CFU of Pseudomonas aeruginosa or PBS as control by intranasal instillation and probed after 12 h. a–b) Pneumonia severity (Mouse Clinical Assessment Scoring System for Sepsis (M-CASS)) and body temperature of the infected and control mice were evaluated. c) Whole-protein analysis of bronchoalveolar lavage (BAL) fluid by bicinchoninic acid kit. d) Wet/dry ratio determination of lung right auxiliary lobes. e) 5-µm section of paraffin embedded lung left lobes were stained with haematoxylin-eosin (H&E), and the thickness of interalveolar septa was determined by blinded analysis of 10 randomly chosen regions of interest for each section of each mouse (representative images are shown). Scale bars=20 μm. Quantitative data are presented as mean±sd (n=6). ns: nonsignificant. *: p<0.05, ***: p<0.001, ****: p<0.0001, two-way ANOVA and Tukey post-test.

FIGURE 5.

Appropriate littermates, Vav-Adam10^−/− or Vav-Adam17^−/− mice were infected with 10⁵ CFU of Pseudomonas aeruginosa or PBS as control by intranasal instillation and probed after 12 h. a, b) Citrated blood, c, d) bronchoalveolar lavage (BAL) fluid or e, f) lung tissue homogenates were serially diluted in sterile PBS and streaked out on LB Agar plates followed by overnight incubation at 37°C to determine P. aeruginosa CFU numbers. Quantitative data are presented as mean±sd (n=6). ns: nonsignificant. **: p<0.01, ***: p<0.001, ****: p<0.0001, one-way ANOVA and Tukey post-test.

FIGURE 6.

Appropriate littermates, Vav-Adam10^−/− or Vav-Adam17^−/− mice were infected with 10⁵ CFU of Pseudomonas aeruginosa or PBS as control by intranasal instillation and probed after 12 h. a, b) Bronchoalveolar lavage (BAL) fluid and c, d) blood were stained with the respective antibodies as described in the methodology section and subjected to flow cytometry to analyse the content of neutrophils, macrophages and monocytes. Quantitative data are presented as mean±sd (n=6). ns: nonsignificant. *: p<0.05, **: p<0.01, ***: p<0.001, two-way ANOVA and Tukey post-test.

ADAM10 changes neutrophil and monocyte properties regulating bacterial clearance in an autocrine and paracrine cell-intrinsic manner

To explain the discrepancy between leukocyte cell numbers and clearance in Vav-Adam10^−/− mice, we investigated the key functional and cellular properties of these cells using primary human neutrophils and THP-1 cells. Pharmacological inhibition of ADAM10 (GI254023X) increased the phagocytosis of GFP-labelled P. aeruginosa as well as the concomitant survival of both cell types significantly, while ADAM17 inhibition (GW280264X) had no effect (figure 7a,b). ROS production was significantly decreased by ADAM10 and increased by ADAM17 inhibition (figure 7c). p38 and Src kinase have been described to be involved in the reciprocal regulation of ROS production and protease activity [30]. Neither inhibition of p38 by SB203580 nor inhibition of Src kinase by PP2 showed any effect on phagocytosis or cell survival. However, inhibition of Scr kinase reduced the generation of ROS significantly (supplementary figure S6a–c). To investigate whether the p38 pathway is generally activated through P. aeruginosa infection, we tested for phosphorylation and total expression of p38. However, no changes were observed (supplementary figure S6d). Differences in the clearing capacity may also result from changes in the killing efficacy. Indeed, pharmacological inhibition of ADAM10, but not ADAM17 increased the bacterial killing capacity of the infected leukocytes measured as bacterial survival (figure 7d). To investigate if the differences in phagocytosis and ROS production might be based on the release of soluble ectodomains (autocrine and paracrine manner) or cell-intrinsic functions, media transfer experiments were performed. The soluble factors released from P. aeruginosa-infected cells (without inhibition) were added to GI254023X or GW280264X pre-incubated cells. The observed increase in phagocytosis by ADAM10 inhibition was abrogated by the transfer of the soluble factors, whereas the differences in ROS generation were not affected (figure 7e,f). Prominent substrates and inducers of ADAM10 and ADAM17 are TNF-α and epidermal growth factor receptor (EGFR) ligands. The blockage of TNF-α-dependent pathways by infliximab reduced the ROS production comparable to ADAM10 inhibition, whereas the inhibition of EGFR signalling by cetuximab had no effect (supplementary figure S6e). These findings strongly suggest that ADAM10 regulates the antimicrobial properties of neutrophils and monocytes both in an autocrine/paracrine (phagocytosis and survival) and cell-intrinsic (ROS production) manner.

FIGURE 7.

a, b) 10⁶ THP-1 cells or human neutrophils were pre-incubated with 10 µM GI254023X, 10 µM GW280264X or 0.1% dimethyl sulfoxide (DMSO) as control for 30 min followed by infection with green fluorescent protein (GFP)-labelled Pseudomonas aeruginosa (PA01) for 2 h. The GFP signals of the nonphagocytosed bacteria were quenched by Trypan blue. Phagocytosis was evaluated by measuring the mean of the GFP signal, and cell survival was evaluated by measuring the mean of the allophycocyanin signal of Trypan blue using flow cytometry (n=6). c) 10⁵ cells of THP-1 or human neutrophils were stained with reactive oxygen species (ROS) red dye for 30 min followed by pre-incubation with 10 µM GI254023X, 10 µM GW280264X or 0.1% DMSO as control for 30 min. Consequently, the cells were infected with P. aeruginosa (PA103, multiplicity of infection 5) for 2 h and the fluorescence signals were monitored at excitation/emission of 520/605 nm (n=4). d) THP-1 cells and neutrophils were infected with GFP-labelled P. aeruginosa (PA01) for 2 h and killing capacity was evaluated by CFU determination and shown as bacterial survival. e, f) Cells were infected with GFP-labelled P. aeruginosa (PA01) for 2 h and media were transferred to GI254023X or GW280264X inhibited cells and infected again with GFP-labelled P. aeruginosa (PA01) for 2 h. Phagocytosis and ROS production were evaluated as described in a) and c). Quantitative data are presented as mean±sd. ns: nonsignificant. *: p<0.05, **: p<0.01, ***: p<0.001, one-sample t-test.

ADAM10-deficiency shapes the local immune response to an anti-inflammatory phenotype

Myeloperoxidase (MPO) is one key enzyme of neutrophils, generating hydrochlorous acid as tissue-damaging ROS. Although Vav-Adam10^−/− mice displayed equal neutrophil numbers in the alveolar space (figure 5a), the MPO activity was strongly reduced compared to that seen in infected littermate mice (figure 8a). Surprisingly, Vav-Adam17^−/−mice showed a strong increase in MPO activity (figure 8b), whereas the cell numbers were reduced (figure 5b). Besides neutrophils, differential activation and polarisation of macrophages has a strong impact on tissue damage. Lung tissue sections were stained against CD68 (general macrophages), CD86 (M1 macrophages) and CD163 (M2 macrophages). The numbers of CD68⁺ and CD86⁺ macrophages were similar between infected Vav-Adam10^−/− and littermate mice (figure 8c,d). However, Vav-Adam10^−/− mice showed a significant increase in CD163⁺ macrophages compared to littermate mice (figure 8e). In contrast, Vav-Adam17^−/− had no impact on either CD68⁺, CD86⁺ or CD163⁺ macrophages (supplementary figure S7). Thus, enhanced clearance and reduced tissue damage in Vav-Adam10^−/− mice may result from enhanced phagocytic properties of leukocytes and reduction of the pro-inflammatory response.

FIGURE 8.

Appropriate littermates, Vav-Adam10^−/− or Vav-Adam17^−/− mice were infected with 10⁵ CFU of Pseudomonas aeruginosa or PBS as control by intranasal instillation and probed after 12 h. a, b) Determination of the bronchoalveolar lavage (BAL) content of myeloperoxidase (MPO) measured using ELISA. c–e) 5-µm sections of paraffin embedded lung left lobes were stained against CD68 (general macrophages), CD86 (M1) and CD163 (M2) to analyse the polarisation of lung macrophages (representative images are shown; cell number per mm² of each section). Scale bars=20 μm. Quantitative data are presented as mean±sd (n=6). ns: nonsignificant. ***: p<0.001, ****: p<0.0001, two-way (a–b) and one-way (c–e) ANOVA and Tukey post-test.

Discussion

Leukocytic ADAM10 and ADAM17 may constitute relevant regulatory elements in bacterial and viral pneumonia through changes of the local and systemic immune response. In the lung microenvironment, leukocytic ADAM10 orchestrated a pro-inflammatory response with M1 polarisation, increased ROS production, cytokine release and tissue damage, whereas ADAM17 seemed to dampen the initial inflammatory response to an anti-infective and protective level. Furthermore, both proteases were systemically released and active on exosomes, at least in vitro leading to pathophysiological changes of cardiomyocytes and the endothelial barrier. In the in vivo situation, this may affect distinct organ sites and modulate the grade of systemic inflammation and disease outcome.

The outcome of pneumonia is strongly dependent on bacterial and viral clearance, as well as the maintenance of barrier integrity. Indeed, Adam10 deficiency increased the phagocyting capacity and the cell survival, both together increasing the bacterial clearance capacity, although a general reduction of migration is known [29]. Treatment evaluations in a murine model of polymicrobial sepsis revealed a beneficial impact of ADAM17 inhibition [31]. In our case, Adam17 deficiency led to detrimental bacterial burden and severe disease development, which can be in parts explained by the reduced diapedesis through, for example, impaired cleavage of l-selectin [32, 33]. Furthermore, ROS production was strongly increased. Under homeostasis, ROS-producing factors are in equilibrium, achieving a redox balance [34]. If this balance is impaired as seen in some COVID-19 patients, the excessive oxidative stress may lead to alveolar damage and thrombosis, resulting in life-threatening circumstances [35] as observed in Vav-Adam17^−/− mice. In Staphylococcus aureus induced pneumonia [36], a Gram-positive bacterium, the benefit of Adam10-deficiency was related to reduced release of IL-1β, but without changes in bacterial clearance or protein leakage. In contrast to these previous reports, we observed enhanced bacterial clearance and reduced protein leakage, indicating reduced tissue damage in Vav-Adam10^−/−mice. S. aureus α-haemolysin uses ADAM10 as pore-forming receptor [37], whereas P. aeruginosa exotoxin A induces a strong calcium influx, activating ADAM10 [15]. Thus, the obvious heterogeneity in cell surface composition and toxin repertoire may be indicative for the differential local ADAM-associated effects [14, 15].

An important function of both ADAM10 and ADAM17 is their stimulus-dependent contribution to IL-6R cleavage resulting in the generation of the sIL-6R [38–40], with Adam17 hypomorphic mice exhibiting unaltered levels of serum sIL-6R [40]. Indeed, serum levels of sIL-6R were not affected in Vav-Adam17^−/− mice, but decreased in Vav-Adam10^−/− mice. IL-6 signalling was shown to be protective in obesity-associated insulin resistance and endotoxaemia through promotion of alternative macrophage activation [12]. Remaining IL-6R on leukocytes due to absence of ADAM10 would allow for efficient IL-6 signalling, thereby promoting M2 polarisation as observed in Vav-Adam10^−/− mice, further explaining the low cytokine levels. Furthermore, reduced levels of serum sIL-6R would dampen the trans-signalling in distinct organs, further protective action as observed for pancreatitis-induced lung injury [41]. In the case of Vav-Adam17^−/− mice, IL-6 signalling may be not affected due to the observed similar levels, together with the reduced clearance capacity and increased ROS production leading to the detrimental effects observed in our study. One important cytokine released by M1 polarised macrophages is TNF-α, further triggering Scr kinase dependent processes such as ROS production or MMP9 release [42]. Indeed, TNF-α neutralisation and Src kinase inhibition, respectively, reduced the ROS production to similar levels than observed for ADAM10 inhibition. Moreover, inhibition of ADAM10 reduced ROS production and macrophages efferocytosis [43]. Thus, ADAM10 seems to interfere with a complex network of inflammatory pathway which may all contribute to its immunomodulatory functions.

A similar complexity in ADAM10 and ADAM17 function has been observed for their direct role in SARS-CoV-2 infection and viral entry. Early in the pandemic phase, anti-TNF-α and anti-IL6R therapies (e.g. infliximab and tocilizumab) were evaluated to prevent severe disease development. Most of the studies did not reach the required end-points, which is why only tocilizumab in combination with corticosteroids is recommended in the current guidelines for severe COVID-19 cases [44]. ACE2 has long been regarded as a substrate of ADAM17 [17]; however, more recent studies indicate a cleavage by both ADAM10 and ADAM17 [18]. It could be argued that an increase of ADAM10/17 activity would be beneficial to prevent viral entry through ACE2, although potentially increasing the amount of other inflammatory mediators such as TNF-α. However, the amount of soluble ACE2 correlated with the mortality in hospitalised COVID-19 patients [45], and the downregulation from the surface worsened the outcome as its protective functions in the cardiovascular system were disabled [46]. More recent studies showed that ADAM17 is unlikely to facilitate SARS-CoV-2 cell entry through cleavage of ACE2, but rather through its interaction and cleavage of the SARS-CoV-2 spike protein, and that ADAM10 fosters the spike protein-mediated syncytia formation at later time points [16]. Indeed, time-dependent processes may challenge the development of ADAM-based anti-infective therapies.

One further level of complexity, potentially reinforcing the systemic inflammatory condition may be the release of exosomes. While in early disease (murine pneumonia model and in vitro infection experiments) active ADAM10 carrying exosomes are increased, ADAM17 in manifested pneumonia (hospitalised patients) gets more prominent. ADAM10 and ADAM17 may be released as soluble variants [47, 48]; however, soluble ectodomain activity could not be observed in the present study. Exosomes as cargo vehicles are generated via endocytosis, maturing to multivesicular bodies, intraluminal budding and final release to the extracellular space [49]. Therefore, they express proteases on the surface in the right orientation [48], retaining their full activity and inducing proteolysis of their substrates in the extracellular space [15, 22]. Protein complexes carrying both agonists, antagonists and intracellular adaptor molecules may be translocated to exosomes [50], including tetraspanins (Tspans) and the inactive rhomboids (iRHOM)1 and 2. Besides, a cargo exchange may occur in the Golgi and endoplasmic reticulum [51, 52], resulting in the inclusion of further intracellular cargos and potentially explaining the presence of pro-ADAM17 on exosomes. iRHOM2 has been described to allow for a dynamic adjustment of the pro-domain amount upon activation, enabling for tight control of activity [53, 54]. However, the dynamic states of mature ADAM17 would result in a transient state in which release of ADAM17 on exosomes would also be possible without a direct involvement of iRHOM2. Tspans, especially members of the Tspan8 subfamily, are extrinsic factors regulating the cleavage activity of ADAM10 distinct from selective binding pockets on the protease itself [55], in parts also regulating the activity of ADAM17 [56]. CD9 and CD81 are markers of exosomes and have been shown to be involved in the cellular trafficking of ADAM10 [57], which could further explain the increased proportion of ADAM10 released on exosomes upon P. aeruginosa infection, while the proportion of ADAM17 stayed constant and was more likely to be increased due to a higher exosome number. Two scenarios are possible. The co-translocation of the proteases and their adaptor molecules might result in a similar tight control of their activity on exosomes, or exosomal ADAM proteases may not need additional activation molecules as they were translocated in the active forms, potentially resulting in a less/uncontrolled cleavage capacity.

At least 50% (basal levels) to 100% (stimulated release) of these exosomes may be derived from leukocytes. However, there might be other sources such as endothelial cells, or platelets contributing to the pool of circulating exosomes. Exosomes trigger pathological responses in a long-distance communication manner [20, 58], and studies on COVID-19 have highlighted the relevance of exosomes for triggering the adaptive immune response and potential induction of multi-organ failure [59, 60]. Exosomes may interact with their recipient cells via close-proximity events and direct interactions, or they may deliver their cargos through receptor-mediated endocytosis, micropinocytosis or membrane fusion events. Extracellular vesicles harbour inflammatory receptors, which may be transferred to the recipient cells, rendering these cells susceptible to inflammatory stimuli [50]. Action of exosomes without fusion have been reported for dendritic cells and B-cell derived extracellular vesicles in T-lymphocyte stimulation and antigen presentation [61, 62]. In the investigated time frame of 2 h, fusion events may be less likely due to a size-related increase in membrane tension and hemifusion stalk energy of exosomes in comparison to extracellular vesicles [63]. We used an inhibition approach ensuring inhibition on the exosomes themselves and not the recipient cells, evidencing a direct functional implication of ADAM harbouring exosomes on endothelial and cardiomyocyte function at least in vitro. Furthermore, we could identify mechanistic substrates cleaved by exosome-associated ADAM10 and ADAM17, including cell-adhesion molecules and VE-cadherin. Excessive cleavage of VE-cadherin and other junction molecules may disrupt structural properties required for endothelial cell survival [64], further evidenced by the increase of apoptosis-related proteins. In addition, exosomes from pneumonia patients increased the presence of PAI-2, which is known to inhibit venous thrombus resolution [65]. COVID-19 increased the risked of thromboembolic events, especially in patients requiring intensive care [66].

Of course, systemic inflammation is driven by locally produced inflammatory mediators disseminated throughout the body, raising the question might ADAM10/17-loaded exosomes serve primarily as biomarker rather than active disease modifier. Based on our data and considering their potential paracrine and endocrine activity, we propose that proteolytically active ADAM carrying exosomes may constitute relevant regulatory elements of long-rang inflammatory signalling, amplifying systemic effects and possibly undermining standard therapy. Further translational and preclinical studies are needed to evaluate the therapeutic potential of ADAM proteases through its immunomodulatory functions and to clarify the diagnostic, prognostic and therapeutic relevance of ADAM10/17-loaded exosomes in standardised, multimodal models.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

ERJ-02060-2024.Supplement