Organizing empyema induced in mice by Streptococcus pneumoniae: effects of plasminogen activator inhibitor-1 deficiency

Abstract

Background

Pleural infection affects about 65,000 patients annually in the US and UK. In this and other forms of pleural injury, mesothelial cells (PMCs) undergo a process called mesothelial (Meso) mesenchymal transition (MT), by which PMCs acquire a profibrogenic phenotype with increased expression of [alpha]-smooth muscle actin ([alpha]-SMA) and matrix proteins. MesoMT thereby contributes to pleural organization with fibrosis and lung restriction. Current murine empyema models are characterized by early mortality, limiting analysis of the pathogenesis of pleural organization and mechanisms that promote MesoMT after infection.

Methods

A new murine empyema model was generated in C57BL/6 J mice by intrapleural delivery of Streptococcus pneumoniae (D39, 3 × 10⁷-5 × 10⁹ cfu) to enable use of genetically manipulated animals. CT-scanning and pulmonary function tests were used to characterize the physiologic consequences of organizing empyema. Histology, immunohistochemistry, and immunofluorescence were used to assess pleural injury. ELISA, cytokine array and western analyses were used to assess pleural fluid mediators and markers of MesoMT in primary PMCs.

Results

Induction of empyema was done through intranasal or intrapleural delivery of S. pneumoniae. Intranasal delivery impaired lung compliance (p < 0.05) and reduced lung volume (p < 0.05) by 7 days, but failed to reliably induce empyema and was characterized by unacceptable mortality. Intrapleural delivery of S. pneumoniae induced empyema by 24 h with lung restriction and development of pleural fibrosis which persisted for up to 14 days. Markers of MesoMT were increased in the visceral pleura of S. pneumoniae infected mice. KC, IL-17A, MIP-1[beta], MCP-1, PGE ₂ and plasmin activity were increased in pleural lavage of infected mice at 7 days. PAI-1^-/- mice died within 4 days, had increased pleural inflammation and higher PGE₂ levels than WT mice. PGE₂ was induced in primary PMCs by uPA and plasmin and induced markers of MesoMT.

Conclusion

To our knowledge, this is the first murine model of subacute, organizing empyema. The model can be used to identify factors that, like PAI-1 deficiency, alter outcomes and dissect their contribution to pleural organization, rind formation and lung restriction.

Tucker et al. Clin Trans Med (2016) 5:17 DOI 10.1186/s40169-016-0097-2

RESEARCH

Open Access

http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = Organizing empyema inducedinmice byStreptococcus pneumoniae: eects ofplasminogen activator inhibitor-1 deciency

http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s40169-016-0097-2&domain=pdf

Web End = Torry A. Tucker*, Ann Jeers, Jake Boren, Brandon Quaid, Shuzi Owens, Kathleen B. Koenig, Yoshikazu Tsukasaki, Galina Florova, Andrey A. Komissarov, Mitsuo Ikebe and Steven Idell

Abstract

Background: Pleural infection aects about 65,000 patients annually in the US and UK. In this and other forms of pleural injury, mesothelial cells (PMCs) undergo a process called mesothelial (Meso) mesenchymal transition (MT), by which PMCs acquire a probrogenic phenotype with increased expression of -smooth muscle actin (-SMA) and matrix proteins. MesoMT thereby contributes to pleural organization with brosis and lung restriction. Current murine empyema models are characterized by early mortality, limiting analysis of the pathogenesis of pleural organization and mechanisms that promote MesoMT after infection.

Methods: A new murine empyema model was generated in C57BL/6 J mice by intrapleural delivery of Streptococcus pneumoniae (D39, 3 1075 109 cfu) to enable use of genetically manipulated animals. CT-scanning and

pulmonary function tests were used to characterize the physiologic consequences of organizing empyema. Histology, immunohistochemistry, and immunouorescence were used to assess pleural injury. ELISA, cytokine array and western analyses were used to assess pleural uid mediators and markers of MesoMT in primary PMCs.

Results: Induction of empyema was done through intranasal or intrapleural delivery of S. pneumoniae. Intranasal delivery impaired lung compliance (p < 0.05) and reduced lung volume (p < 0.05) by 7 days, but failed to reliably induce empyema and was characterized by unacceptable mortality. Intrapleural delivery of S. pneumoniae induced empyema by 24 h with lung restriction and development of pleural brosis which persisted for up to 14 days. Markers of MesoMT were increased in the visceral pleura of S. pneumoniae infected mice. KC, IL-17A, MIP-1, MCP-1, PGE2 and plasmin activity were increased in pleural lavage of infected mice at 7 days. PAI-1/ mice died within 4 days, had increased pleural inammation and higher PGE2 levels than WT mice. PGE2 was induced in primary PMCs by uPA and plasmin and induced markers of MesoMT.

Conclusion: To our knowledge, this is the rst murine model of subacute, organizing empyema. The model can be used to identify factors that, like PAI-1 deciency, alter outcomes and dissect their contribution to pleural organization, rind formation and lung restriction.

Keywords: Pleural mesothelial cells, Pneumonia, Plasminogen activator inhibitor-1

Background

Pleural infection remains a common and important clinical problem that can complicate pneumonia or occur

after trauma. The incidence of empyema continues to rise despite the broad use of vaccinations and development of more potent antibiotics [1, 2]. Of the roughly 4 million cases of pneumonia annually diagnosed in the US, about half develop a parapneumonic eusion [1]. In many cases, these eusions can evolve with formation of complicated parapneumonic eusions or frank empyema, which is characterized by overt infection or intrapleural

*Correspondence: [email protected] Department of Cellular and Molecular Biology, The University of Texas Health Science Center at Tyler, 11937 US HWY 271, Biomedical Research Building, Lab C-5, Tyler, TX 75708, USA

2016 The Author(s). This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Tucker et al. Clin Trans Med (2016) 5:17

Page 2 of 18

pus. Biomarkers such as low pH predict the development of pleural loculation [1]. About 65,000 patients in the US and UK suer from pleural infection each year [1, 3], which is associated with increased morbidity, mortality and medical costs approaching half a billion dollars annually [4, 5]. About 40,000 US patients annually may require pleural drainage to prevent morbidity associated with complicated parapneumonic eusions/empyema in the US each year [6]. Current surgical treatment is invasive and alternative treatment with intrapleural administration of brinolysins is associated with variable outcomes in adults [7]. These considerations justify the search for more eective interventions, which rely on better understanding of the pathogenesis of pleural organization and remodeling [8].

In a previous publication, we showed that a combination of bleomycin and carbon black (CBB) could reliably and reproducibly induce pleural injury including reduced lung function and increased pleural thickening in C57BL/6 mice [9]. We also reported a Pasteurella multocida model of pleural injury in rabbits [10]. Although this injury is quite robust, simulates human pleural injury and has been used to test the efficacy of brinolytic agents, this pathogen rarely causes pleural infection in clinical practice. Because the use of larger animals limits use of transgenic or knock-out animals, a murine model of pleural injury and repair in the C57BL/6 strain is desirable. While a previous study showed that intranasally administered S. pneumoniae (D39 strain) produced intrapleural injury in CD1 mice, that model was characterized by early mortality that restricted analyses to the acute setting; over 48h [11]. Pleural infection was lethal thereafter, precluding analysis of the remodeling that occurred after acute injury. To our knowledge, no murine model of empyema and progressive pleural organization after bacterial infection has been reported, which has slowed progress in the eld. We inferred that a more durable model of infectious pleural injury could be developed and tested that postulate using C57BL/6 mice to develop a model of organizing of S. pneumoniae pleural empyema with survivorship over 2weeks. We used the model to assess the impact of PAI-1 deciency on the evolution of organizing empyema, as this derangement was previously reported to increase pleural rind formation and lung restriction in the CBB model of noninfectious pleural injury [9].

Methods

Intranasal andintrapleural inoculations

All experiments involving animals were approved by the Institutional Animal Care and Use Committee at the University of Texas Health Science Center at Tyler. C57BL/6 J mice (1012 weeks of age, 20 g,

Jackson Laboratory, Bar Harbor ME) were lightly

anesthetized with isourane. Intranasal inoculations (31075109) of Streptococcus pneumoniae (S. pneumoniae, D39, National Collection of Type Cultures,

Salisbury UK) resuspended in 0.9 % saline were delivered in 40 l over the nares. Intrapleural inoculations (5 1075 108 cfu, resuspended in 0.9 % saline) of

S. pneumoniae were delivered in 150L by intrapleural injection. The control group received normal saline under the same conditions. Antibiotic treatment (enrooxacin, 15 mg/kg) was initiated 18 h post infection and was administered daily by subcutaneous injection for 4 days. Mice were periodically monitored following infection to record body weight, dehydration status, activity and behavior. If dehydration was detected by alterations of skin turgor, the aected mice were subcutaneously injected with 200500 l of warmed 0.9% saline, as needed. Moribund animals were euthanized. After administration of S. pneumoniae, mice were housed on a heating pad to maintain an ambient temperature of 30C throughout the time course.

Lung andpleural lavage collection

Lung and pleural lavages were performed using 1.5ml of sterile normal saline were immediately performed at the time of death in selected animals, as previously described [9]. Total white cell and dierential cell counts were likewise measure in these uids, as we previously reported [9].

Cultures ofpleural uids

Pleural lavages of saline and S. pneumoniae infected mice were cultured on blood agar plates containing 5% sheep blood (Remel Blood Agar, Fisher Scientic). Neat (50l) and 1:100 dilutions of the pleural lavages were cultured on blood agar plates and incubated 15h at 37C. Colonies were then counted to determine bacterial burden.

Lung histology, immunostaining, confocal, bright eld microscopy andmorphometry

Lung histology and immunostaining were performed as previously described [9, 12]. All tissue sections were rst deparanized and subjected to antigen retrieval using a citrate buer at 95 C for 20 min. Tissue analyses, collagen deposition and localization were initially assessed by Trichrome staining as previously described [9, 12]. Morphometric analyses of pleural tissue thickness and depth of underlying pneumonitis were performed as we previously described [9]. Fibrin (ogen) antigen was assessed using immunohistochemistry (IHC) and Fast Red (BioGenex, Freemont CA) chromogen as previously described [13].

Immunouorescence was used to visualize -SMA and calretinin expression in saline and S. pneumoniae infected

Tucker et al. Clin Trans Med (2016) 5:17

Page 3 of 18

pleuropulmonary sections as previously described [9]. Confocal microscopy was then used to visualize immunouorescence and co-localization of the markers. Confocal images were acquired from a eld of view at 0.4-m z-axis increments with the LSM 510 Meta confocal system (Carl Zeiss) at 40 as previously described [9, 13].

Collagen detection inlung tissues

Bright eld microscopy was used to image trichrome stained tissue sections as previously described [9]. Collagen was detected by picrosirius staining and imaged using a polarized light source on a Nikon Ti inverted microscope.

Pulmonary function testing

Pulmonary function tests were performed immediately before CT imaging and prior to sacrice, as previously described [9, 12]. Briey, mice were anesthetized with a ketamine/xylazine mixture. Anesthetized mice were intubated by inserting a sterile, 20-gauge intravenous cannula through the vocal cords into the trachea. Measurements were then performed using the exiVent system (SCIREQ, Tempe AZ). The snapshot perturbation method was used to determine lung compliance, according to manufacturers specications. Mice were maintained under anesthesia using isourane during pulmonary function testing.

Computed tomography (CT) scans andmeasurements oflung volume

Chest CT imaging and measurements of lung volume were performed as previously described [9, 12]. Keta-mine/xylazine anesthetized mice were anesthetized further using an isourane/oxygen mixture to minimize spontaneous breaths and to ensure that mice remained anesthetized throughout the procedure. Images were obtained using the Explore Locus Micro-CT Scanner. CT scans were performed during full inspiration and at a resolution of 93m. Microview software was used to analyze lung volumes and render three-dimensional images. Lung volumes were calculated from renditions collected at full inspiration.

Measurement ofpleural uid plasmin andbrinolytic activity

Plasmin activity in the pleural washes of WT and PAI-1/ mice treated with saline or S. pneumoniae was measured by amidolytic assay using a plasmin substrate (PL-5268, Centerchem Inc, Norwalk CT) on a Spec-traMax 96-well optical absorbance plate reader (Molecular devices, Sunnyvale, CA), as previously described [14]. Fibrinolytic activity was measured as previously described [14].

Bioplex analyses ofpleural lavage mediators

Analyses of pleural lavage inammatory mediators including: Eotaxin, G-CSF,GM-CSF,IFN-, IL-1, IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-17A, KC, MCP-1 (MCAF), MIP-1, MIP-1, RANTES, TNF- were determined using the Bio-Plex Pro Mouse Cytokines 23-plex (BIO-RAD) on a BioPlex MAGPIX Multiplex Reader according to the manufacturers instructions.

PGE2 ELISA

PGE2 was rst extracted from pleural lavage by C18 column purication according to manufacturers directions (Cayman Chemical, Ann Arbor Michigan). PGE2

levels were determined by competitive ELISA (Cayman) according to manufacturers instructions.

Primary pleural mesothelial cell culture andtreatment

Permission to collect and use HPMCs was granted through an exempt protocol approved by the Institutional Human Subjects Review Board of the University of Texas Health Science Center at Tyler. HPMCs were isolated from pleural uids collected from patients with congestive heart failure or post-coronary bypass pleural eusions as previously described [15]. HPMCs were maintained in LHC-8 culture media (Life Technologies, Carlsbad CA) containing 3 % fetal bovine serum (Life Technologies), 2 % antibioticantimycotic (Life Technologies) and Glutamax (Life Technologies) as previously described [9, 13, 15, 16]. MPMCs were isolated and cultured as previously reported [9]. All cells were cultured in a humidied incubator at 37C in 5% CO2/95% air. Cells were passaged a maximum of ve times before discontinuing use. Serum-starved cells were treated with TGF- (5ng/ml), PGE2 (1M), butaprost (EP2 agonist, 1M), sulprostone (EP3 agonist, 1 M) and Cay10598 (EP4 agonist, 1M, Cayman). Cell lysates were then Western blotted for -SMA and -actin as previously described [9, 17].

Statistics

All statistics were performed using the MannWhitney U test. A p value of less than 0.05 was considered signicant.

Results

Intranasal administration ofS. pneumoniae did not reliably induce survivable empyema

To create a model of pneumonia complicated by empyema, we initially challenged C57BL/6 mice with intra-nasal administration of S. pneumoniae. A range of doses were used to induce empyema in doseresponse analyses (3 1075 109 cfu). Intranasal administration of

Tucker et al. Clin Trans Med (2016) 5:17

Page 4 of 18

S. pneumoniae (3108 CFU) signicantly reduced lung volume (Fig. 1a, p < 0.01) and decreased compliance (p<0.01). As previously reported [11], intranasal administration of S. pneumoniae was associated with signicant mortality without antibiotics (data not shown), so that a quinolone to which the organism is exquisitely susceptible was begun at 18h. Doses of less than 3108 cfu did not induce detectable changes in lung function. A range of higher intranasal doses (up to 109 CFU) consistently caused early mortality (<24h) associated with severe diffuse pneumonitis, pleural inammation and pleural eusions and were not further investigated (data not shown). Tissue sections from saline and infected mice were next analyzed to assess changes in lung histology and collagen deposition. Alveolar inammation was apparent at 7days after 3108CFU and mild reactive changes of the visceral pleurae were consistently observed (Fig.1b). Pleural thickness was not signicantly increased by intranasal S. pneumoniae infection (data not shown). No pleural eusions or adhesions were found at 7days in any of these animals, indicating that pleural injury was relatively mild

in animals treated with a maximal, survivable intranasal dose of the organism.

Intrapleural administration ofS. pneumoniae induces pleural organization andremodeling withrind formation

Because of the lack of pleural organization with sublethal doses of intranasally administered S. pneumoniae, we next sought to determine if intrapleural administration yielded survivable empyema. S. pneumoniae was therefore injected directly into the pleural space, at a range of doses (51075108 cfu) with no antibiotic coverage. We found that doses of less than 1108 CFU did not reliably induce pleural injury, while doses above 3108 cfu were associated with unacceptable early mortality of almost all animals (data not shown). S. pneumoniae infected mice demonstrated dramatic decrements in weight for the rst 3days after infection (Fig.2a). By 7days, the infected mice had regained most of the lost weight. In initial analyses, mice were euthanized 7days after S. pneumoniae administration. Gross analyses of infected mice showed increased deposition of transitional

Fig. 1 Intranasal S. pneumoniae causes pneumonia, lung dysfunction and mild pleural inammation at survivable intranasal dosing. S. pneumoniae (3 108CFU) was intranasally administered to WT mice and maintained for 7 days. a Lung volumes were measured by CT scan and pulmonary

function (compliance) measured as described in the Methods. n = 67 mice/group. b Peripheral lung tissue sections from saline-challenged and

S. pneumoniae infected mice collected at 7 days were stained with trichrome to show changes in lung architecture and collagen deposition (blue stain). S. pneumoniae infected mice exhibited thickened alveolar septae, mild pleural inammation, reactive visceral mesothelial cells and mildly increased collagen deposition compared to saline controls. No pleural uid or brinous strands at the pleural surface were seen. Solid arrows indicate reactive mesothelial cells. Images are 40 and are representative of the ndings of 30 elds/slide and 67 mice/group

Tucker et al. Clin Trans Med (2016) 5:17

Page 5 of 18

a

b

c

d

75 p<0.001

Pleural Thickness

(m)

50

25

0 WT Saline 7D

WT Strep 7D

Tucker et al. Clin Trans Med (2016) 5:17

Page 6 of 18

(See gure on previous page.)

Fig. 2 Intrapleural S. pneumoniae induces pleural injury and lung dysfunction. S. pneumoniae was intrapleurally administered and incubated for7 days. a Weights were collected over 7 days (n = 37 mice/group). b Gross images taken at 7 days after administration of saline or S. pneumoniae.

These animals did not receive antibiotic treatment. S. pneumoniae infection promotes extensive pleural injury with deposition of a transitional intrapleural brinous neomatrix that is readily appreciated with coating and encasement of the lungs. Solid arrows indicate areas of neomatrix deposition within the thoracic cavity of S. pneumoniae infected mice. c Lung volumes and function (compliance) were signicantly reduced by S. pneumoniae administration at 7 days (p < 0.05). d Lung tissue sections from saline-treated and S. pneumoniae infected mice (7 days post-infection) were trichrome stained, with collagen deposition indicated by the blue stain. By morphometry, S. pneumoniae infected mice demonstrated signicantly more pleural thickening than saline controls (p < 0.001). Solid arrows indicate the pleural surface and the basement membrane of the thickened pleura. Areas of collagen deposition were observed within the thickened pleural rind of S. pneumoniae infected mice. Images were obtained at 20 and are representative of the nding of 30 elds/slide and 37 mice/group

brinous adhesions in the pleural space (Fig.2b) and the pleural surface was coated with what grossly appeared to be purulent material by 2days (data not shown). This injury progressively worsened over 7 days, with pleural thickening detectable by 5 days (not shown) and increased by 7 days, often with areas of discrete pleurodesis. As anticipated based on the gross ndings, intrapleural administration of S. pneumoniae signicantly impaired lung volume (p < 0.05) and lung function; compliance (p < 0.05) by 7 days (Fig. 2c). By histologic analysis, infected mice exhibited extensive remodeling characterized by signicant pleural thickening with collagen deposition (blue stain) at 7days compared to saline-treated controls (Fig.2d). Ultimately, we found that this model was limited by mortality of greater than half of the animals by 7days, which was unacceptable.

Antibiotic treatment improves survival ofmice intrapleurally infected withS. pneumoniae

To expand the experimental time frame so that pleural organization and remodeling could be interrogated over 714 days, intrapleurally infected mice were given an antibiotic; enrooxacin, 18h after intrapleural administration of S. pneumoniae (1.8108). Antibiotic-treated

S. pneumoniae-infected mice demonstrated rapid weight loss by 3 days after infection (Fig. 3a), similar to mice without antibiotic treatment. However, these mice began to gain weight by 4days and approached saline control levels by 6 days. Gross analyses (Fig. 3b) showed pronounced pleural injury in antibiotic-treated, S. pneumoniae infected mice at 7days (panel b) when compared to saline treated mice (panel a).

We next performed pulmonary function tests and CT scans at 7days in these mice and found consistent decrements in lung volume (p<0.01) and pulmonary function; compliance (p < 0.01), versus saline controls (Fig. 3c). Total WBC were signicantly elevated (17.77.8105

versus 8.6 2.3 105 p < 0.01, n = 8) in the pleural lavages of infected mice compared to saline controls at 7days with antibiotic treatment. The median neutrophil percentage was 25% at 7days in the empyema animals

and was 1% in the saline controls. Tissue sections from saline and S. pneumoniae infected mice were next analyzed by histology. Because parietal pleural injury was more heterogeneous and responses of the visceral and parietal pleural surfaces were found to be comparable, we focused on determination of changes in the visceral pleura, as we previously reported [9]. S. pneumoniae infected mice exhibited pronounced matrix deposition and signicant pleural thickening (p < 0.001, Fig. 3d) compared to saline controls. The infection also induced collagen expression in the mesothelium and subpleural mesothelium by 7 days as determined by Picrosirius staining for collagen (redorange birefringence, Fig.4a). Tissue sections from saline and S. pneumoniae-infected mice were next stained for brin(ogen) by immunohistochemistry (red stain, Fig. 4b). Robust brin (ogen) deposition was detected at the visceral pleural surface 7 days post-infection. That both brin(ogen) and collagen are detectable in the thickened visceral pleura by 7days after S. pneumoniae injury indicates that ongoing pleural organization contributed to the restrictive, physiologic alterations we observed at this time. Pleural lavage from euthanized saline and S. pneumoniae infected mice were cultured on blood agar dishes to determine if live S. pneumoniae persisted throughout the 7days time course. With antibiotic treatment, pleural lavages were sterile by 7 days after intrapleural inoculation, while live S. pneumoniae were routinely cultured from the lavages of 7days infected mice that were not treated with antibiotics. In time course experiments, S. pneumoniae colonies could not be detected in pleural lavage by 3 days after antibiotic administration (data not shown).

Because antibiotic treatment increased survival of infected mice, we next extended the model over 14days and determined the eect of S. pneumoniae infection on pulmonary function (Fig. 4c). Signicant decrements in lung volume (p<0.01) and pulmonary function; compliance (p < 0.01) persisted at 14 days. Increased collagen deposition (blue stain) and signicant increases in pleural thickness remained at 14days (Fig.5a). Underlying pneumonitis was detected in S. pneumoniae infected mice at

Tucker et al. Clin Trans Med (2016) 5:17

Page 7 of 18

a

b

c

d

150 p<0.001

Pleural Thickness

(m)

100

50

0 WT Saline 7D WT Strep 7D

Fig. 3 Antibiotic treatment extends survival after S. pneumoniae-induced empyema which is characterized by persistent pleural brosis, rind formation and lung restriction. Antibiotic treatment was begun 18 h after S. pneumoniae was intrapleurally administered. a Weights were collected overa 7 days time course (n = 8 animals/group). The infected mice recovered lost weight by 6 days. b Gross images were taken 7 days after intrapleural

administration of saline or S. pneumoniae. S. pneumoniae infection promotes progressive pleural injury and deposition of a brinous matrix with antibiotic treatment (subpanel B) when compared to saline controls (subpanel A). Representative images are shown (n = 68 animals/group). c Lung

volumes and function (compliance) were signicantly reduced by S. pneumoniae infection by 7 days (P < 0.01). d Lung tissue sections from saline and S. pneumoniae infected mice were trichrome stained for collagen deposition (blue). Solid arrows indicate the pleural surface and the basement membrane of the thickened pleura. Areas of collagen deposition were observed within the thickened pleural rind of S. pneumoniae infected mice. S. pneumoniae-infected mice had signicantly increased pleural thickening when compared to saline controls (68 mice/group)

Tucker et al. Clin Trans Med (2016) 5:17

Page 8 of 18

Fig. 4 S. pneumoniae-induced empyema is characterized by persistent pleural brosis, rind formation and lung restriction at 7 days. Antibiotic treatment was begun 18 h after S. pneumoniae was intrapleurally administered. a Peripheral lung sections from saline and S. pneumoniae infected mice were picrosirius stained to detect collagen deposition (redorange birefringence). S. pneumoniae infected mice demonstrated increased collagen deposition at the pleural surface and in the subpleural region compared to saline-treated mice. Solid arrows indicate the basement membrane and the visceral pleural surface. Areas of collagen deposition were detected at the basement membrane and within the thickened pleural mesothelium. b Tissue sections from 7 days saline and S. pneumoniae infected mice were stained for brin (ogen) by immunohistochemical analysis (red stain). Extravascular brin (ogen) deposition is readily apparent after S. pneumoniae infection. Solid arrows indicate the pleural surface and the basement membrane of the thickened pleura of S. pneumoniae infected mice. Fibrin(ogen) deposition was prominent at the pleural surface of S. pneumoniae infected mice. c Lung volumes and function (compliance) were signicantly decreased by S. pneumoniae infection by 14 days (p < 0.01)

14 days mice. Pleural brin deposition and intrapleural adhesions were likewise detectable at this time (data not shown). Collagen deposition, by picrosirius staining, persisted in 14days infected WT mice compared to saline controls (Fig.5b). Pleural thickening did not signicantly change between 7 and 14days.

Mediator Prole inpleural uids ofS. pneumoniae infected mice

We next sought to characterize the prole of locally elaborated mediators promoting signicant increases in WBC that were observed in pleural uids of infected mice. We rst measured changes in inammatory mediators. KC,

Tucker et al. Clin Trans Med (2016) 5:17

Page 9 of 18

a

75

Pleural Thickness

(m)

p<0.0001

50

25

0 WT Saline 14D WT Strep 14D

b

Fig. 5 S. pneumoniae-induced empyema is characterized by persistent pleural brosis, rind formation and lung restriction at 14 days. Antibiotic treatment was begun 18 h after S. pneumoniae was intrapleurally administered. a Lung tissue sections from saline and S. pneumoniae-treated mice were Trichrome stained to assess collagen deposition (blue) and determine pleural thickness. 14 days S. pneumoniae-infected mice exhibited signicantly increased pleural thickening compared to saline controls. Solid arrows indicate the pleural surface and the basement membrane of the thickened visceral pleura. Broken arrows indicate areas of pneumonitis underlying the thickened pleura. Collagen deposition was detected within the pleural mesothelium of S. pneumoniae infected mice. Images were collected at 20 and are representative of 30 elds/slide and 68 mice/group.

b Lung tissue sections from saline and S. pneumoniae treated mice were stained with picrosirius to conrm tissue collagen deposition (redorange birefringence). S. pneumoniae infected mice demonstrated increased collagen deposition at the pleural and subpleural space when comparedto saline-treated mice at 14 days. Solid arrows indicate areas of collagen deposition within the mesothelium of the visceral pleura of saline and S. pneumoniae infected mice. Images were taken at 20 and are representative of 30 elds/slide and 68 mice/group

IL-17A, MIP-1 and MCP-1 were signicantly increased with S. pneumoniae infection by 7 days (p < 0.05) but were not signicantly dierent from control levels by 14days (Fig.6a). Conversely, IL-13, an anti-inammatory cytokine was signicantly down-regulated (p < 0.01) in S. pneumoniae-infected mice by 7 days but returned to

saline controls levels by 14 days. Eotaxin, G-CSF,GMCSF,IFN-, IL-1, IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 (p40), IL-12 (p70), MIP-1, RANTES, TNF- were not signicantly changed at 7 and 14 days post infection. PGE2, an inammatory cytokine found to be increased in pleural injury [18], was likewise increased in

Tucker et al. Clin Trans Med (2016) 5:17

Page 10 of 18

a

1500 p<0.01

400

750

p<0.05

p<0.01

300

1000

500

KC

(pg/mL)

IL-17A

(pg/mL)

200

MIP-1~

(pg/mL)

500

250

100

0

0

0

7d Saline

7d Strep

14d Saline

14d Strep

7d Saline

7d Strep

14d Saline

14d Strep

7d Saline

7d Strep

14d Saline

14d Strep

3000

800

p<0.01

p<0.01

600

2000

MCP-1

(pg/mL)

IL-13

(pg/mL)

400

1000

200

0

0

7d Saline

7d Strep

14d Saline

14d Strep

7d Saline

7d Strep

14d Saline

14d Strep

*

b

300

200

PGE 2

(pg/mL)

100

0

7d Saline 7d Strep

Fig. 6 Inammatory cytokine prole of S. pneumoniae infected mice. Lavages were isolated from the pleural cavities of saline and intrapleurally infected S. pneumoniae infected mice 7 and 14 days post infection. a Proinammatory mediator expression was then assayed in the collected lav-ages by Bioplex assay. Pleural lavages showed signicant changes in KC, IL-17A, MCP-1, MIP-1, and IL-13 at 7 days by Bioplex analyses (P < 0.01).n = 68 mice/group. PGE2 was extracted from the PLs of saline and S. pneumoniae-infected mice. PGE2 levels were then determined by competitive

ELISA. b PGE2 expression was signicantly increased in the lavages of S. pneumoniae infected WT mice at 7 days (*denotes a P < 0.05), n = 68 mice/

group

these pleural lavage samples versus saline control mice at 7days post-induction of empyema (Fig.6b).

pneumoniaemediated pleural injury is characterized byprominent visceral pleural MesoMT

We previously showed that MesoMT of visceral pleural mesothelial cells likely contributes to the formation of the pleural rind associated with non-specic pleuritis in human tissues [9]. The contribution of MesoMT to pleural rind formation was also conrmed in carbon black/

bleomycin-mediated pleural injury [9]. We next sought to determine the contribution of MesoMT to pleural remodeling in S. pneumoniae-mediated pleural injury using confocal microscopy. Saline treated mice were positive for the mesothelial protein; calretinin (green) but did not express the myobroblast marker; -SMA (red, Fig. 7) in the mesothelium and submesothelial tissues. Conversely, -SMA expression was enhanced at the pleural surface and colocalized with calretinin in S. pneumoniae-infected mice. These ndings strongly suggest that

Tucker et al. Clin Trans Med (2016) 5:17

Page 11 of 18

Fig. 7 Mice with S. pneumoniae empyema demonstrate increased -SMA in the mesothelium of the visceral pleura. Peripheral lung tissue sections from saline and S. pneumoniae-infected mice were prepared from the lungs of mice with empyema that progressed over 7 days. The sections were then probed for calretinin (green) and -SMA (red) expression in the pleural mesothelium and observed by confocal microscopy. Solid arrows indicate areas of -SMA and calretinin colocalization at the surface and within the thickened visceral pleura of S. pneumoniae infected mice. All images were taken at 40 magnication. Images are representative of 30 elds/slide and n = 56 mice/group

MesoMT of resident pleural mesothelial cells localizes to and contributes to the extensive pleural rind formation and remodeling associated with S. pneumoniae-induced pleural organization.

PAI1 deciency worsens S. pneumoniae pleural injury at3days postinfection

We recently showed that PAI-1 deciency exacerbated carbon black/bleomycin-mediated pleural injury [9]. To interrogate the role of PAI-1 in the progression of S. pneumoniae empyema, PAI-1/ mice were used. Unlike WT mice, infected PAI-1/ mice that received antibiotic treatment demonstrated signicant mortality by 4 days (data not shown) so that we could only reliably maintain these animals over 3days. Although these animals only received 2 days of the antibiotic regimen, no bacteria were detected in the pleural washes of either WT or PAI-1 infected mice at 3days (data not shown). While infected PAI-1/ mice demonstrated signicant changes in lung volume (p = 0.02) by 3 days (Fig. 8a), trends in infected WT mice did not reach signicance by 3 days. S. pneumoniae-infected WT and PAI-1/ mice exhibited signicant changes in compliance when compared to the saline controls (p=0.03 versus 0.01 respectively). Total WBC counts in pleural lavage were also signicantly increased in S. pneumoniae infected WT

(1.160.24106 versus 11.514.7106, p<0.05) and PAI-1/ (1.460.80106versus 12.73.01106, n=46 mice/group, p<0.05) versus saline controls. The percentage of lavage neutrophils was also signicantly increased in PAI-1/ decient mice (p<0.0001, Fig.8b).

Tissue sections from S. pneumoniae-infected WT and PAI-1/ mice were next stained for brin (ogen) by IHC (Fig.8c). S. pneumoniae infected PAI-1/ mice demonstrated decreased pleural brin deposition that was readily apparent compared to WT mice treated in the same manner. Extravascular brin (ogen) was not detected in saline treated mice (data not shown). Tissues from 3days infected WT and PAI-1/ mice were also stained for -SMA to determine the extent of MesoMT. However there were no dierences between infected WT and PAI-1/ mice at 3days (data not shown).

We next determined the inammatory mediator prole in pleural lavage of 3days WT and PAI-1/ infected mice. WT and PAI-1/ S. pneumoniae infected mice demonstrated signicant increases in KC, IL-4, IL-5, IL-6, IL-10, IL-12, MCP-1, G-CSF and MIP-1 (Fig. 9, P < 0.05) compared to the saline controls. However there were no signicant dierences between WT and PAI-1/ mice, saline or infected. Conversely, IL-13 expression was signicantly decreased in WT mice with infected with S. pneumoniae (p<0.05) at 3days. While

Tucker et al. Clin Trans Med (2016) 5:17

Page 12 of 18

a

0.08

700

p=0.02

Lung Volume

Lung Compliance

(mL/cm H 2O)

ns

p=0.03 p=0.01

600

0.06

(mm3 )

500

400

0.04

300

0.02

WT Saline n = 4

WT

Strep n = 5

Saline n = 4

-/

PAI-1

Strep n = 7

-/

PAI-1

WT Saline n = 4

WT Strep n = 5

Saline n = 4

Strep n = 7

-/-

-/- PAI-1

Neutrophils

b

75 p<0.0001

% of Total WBC

50

25

0

WT PAI-1-/-

c

Fig. 8 Pleural inammation and lavage neutrophilia is increased in PAI-1 deciency after induction of S. pneumoniae mediated pleural injury. Pleural injury was induced in WT and PAI-1/ mice by intrapleural injection of S. pneumoniae and maintained for 3 days. CT scan and pulmonary function analyses were performed 3 days after infection. a S. pneumoniae infection induced signicant pleural injury in PAI-1/ mice (volume and compliance, p = 0.02 and p = 0.01, respectively). b Pleural lavage uid from S. pneumoniae infected PAI-1/ mice contained a signicantly higher

percentage of neutrophils than identically treated WT mice. Neutrophil percentage was determined by dierential staining and cellular dierential performed at 100 oil magnication, n = 47/group. c Tissue sections from saline and S. pneumoniae-infected mice were prepared from the lungs

of WT and PAI-1/ mice after 3 days. The sections were then stained for brin (ogen) (red) by IHC. All images were taken at 20 magnication.

Images are representative of 30 elds/slide and n = 47 mice/group. Pleural lavages (PL) were performed on S. pneumoniae infected mice at 3 days

Tucker et al. Clin Trans Med (2016) 5:17

Page 13 of 18

KC

IL-6

2500

3000

*

2000

* *

KC

(pg/mL)

1500

2000

IL-6

(pg/mL)

1000

1000

*

500

0

0

3d Saline

3d Strep

Saline

3d Saline

3d Strep

Saline

3d PAI-1

Strep

-/-

-/- d PAI-1

3d PAI-1

Strep

-/-

-/- 3d PAI-1

3

IL-4

IL-5

100

*

* *

450

75

360

*

IL-4

(pg/mL)

270

50

IL-5

(pg/mL)

180

25

90

0

0

3d Saline

3d

Strep

Saline

3d Saline

3d Strep

Saline

3d PAI-1

Strep

-/-

-/- 3d PAI-1

3d PAI-1

Strep

-/

-/- 3d PAI-1

IL-10

IL-12 (p40)

1200

*

750

* *

900

IL-10

(pg/mL)

*

500

600

IL-12 (p40)

(pg/mL)

250

300

0

0 Saline

3d Strep

Saline

Strep

-/-

3d Saline

3d

Strep

Saline

Strep

-/-

3d

PAI-1

-/- 3d PAI-1

PAI-1

-/- 3d PAI-1

3d

3d

*

MCP-1

G-CSF

12000

5000

*

*

3750

MCP-1

(pg/mL)

G-CSF

(pg/mL)

8000

2500

4000

1250

*

0

0

3d Saline

3d Strep

Saline

3d Saline

3d Strep

Saline

3d PAI-1

Strep

-/-

-/- 3d PAI-1

3d PAI-1

Strep

-/-

-/- 3d PAI-1

MIP-1~

*

2500

800

*

2000

* *

600

MIP-1~

(pg/mL)

ns

1500

IL-13

(pg/mL)

400

1000

500

200

0

0

3d Saline

3d Strep

Saline

3d Saline

3d Strep

Saline

3d PAI-1

Strep

-/-

-/- 3d PAI-1

3d PAI-1

Strep

-/-

-/- 3d PAI-1

Fig. 9 Pleural inammation and lavage neutrophilia is increased in PAI-1 deciency after induction of S. pneumoniae mediated pleural injury. Pleural injury was induced in WT and PAI-1/ mice by intrapleural injection of S. pneumoniae for 3 days. PLs from saline and S. pneumoniae-infected WT and PAI-1/ mice demonstrated signicant increases in KC, IL-4, IL-5, IL-6, IL-10, IL-12, MCP-1, G-CSF, and MIP-1 * denotes P < 0.05, n = 47 mice/

group). S. pneumoniae infection signicantly decreased IL-13 in WT mice (p < 0.05) when compared to controls. Saline-treated PAI-1/ mice demonstrated signicantly less IL-13 than similarly treated WT mice (p < 0.05). n = 47 mice/group

Tucker et al. Clin Trans Med (2016) 5:17

Page 14 of 18

IL-13 levels in saline treated PAI-1/ mice were not signicantly dierent from S. pneumoniae infected PAI-1/

mice, they were signicantly lower than saline treated WT mice (p<0.05).

PGE2 andplasmin are increased inPAI1/ mice withempyema

Because PGE2 was signicantly increased in the pleural lavages of 7 days S. pneumoniae-infected WT mice, we next assayed PGE2 levels in the pleural lavages of 3day infected WT and PAI-1/ mice. PGE2 levels were significantly increased in the pleural lavages of S. pneumoniae-infected PAI-1 decient mice compared to identically treated WT mice at 3days (p<0.05, Fig.10a). Because plasmin has been reported to increase COX-2 and, consequently, PGE2 expression [19], we next assayed plasmin activity in the pleural lavages of WT and PAI-1/ mice (Fig. 10b). As anticipated, plasmin activity was signicantly increased in the pleural lavage of S. pneumoniae-infected WT and PAI-1/ mice compared to saline controls (P = 0.05 and 0.02 respectively). Further, S.

pneumoniae infected-PAI-1/ mice demonstrated signicantly higher plasmin activity than infected WT mice (p = 0.04). These ndings were conrmed by measuring the brinolytic potential of these uids using FITC-labeled brin as previously described (data not shown) [14].

PGE2 induces MesoMT

Our data suggest that increased plasmin activity in the pleural uids of S. pneumoniae infected PAI-1 decient mice may contribute to induction of PGE2. uPA is also increased in pleural lavage of injured PAI-1/mice [9]. Therefore, we next tested the ability of uPA and plasmin to induce cyclooxygenase-2 (COX-2) and PGE2 in human (H) and murine (M) PMCs. uPA- and plasmin induced COX-2 and signicantly increased PGE2 expression by

HPMCs (Fig.11a, b, respectively) and MPMCs (Fig.11c, d, respectively). Because PGE2 was increased in pleural injury and increased PGE2 levels occurred in pleural lavage of PAI-1/ mice, we next determined the ability of PGE2 to induce biomarkers of MesoMT using PMCs.

PGE2 induced -SMA protein expression in HPMCs and MPMCs (Fig.12a, b respectively). TGF- was used as a positive control in both HPMCs and MPMCs as it is an established stimulus of biomarkers of MesoMT; -SMA in HPMCs [9, 17]. Phenotypic changes indicative of MesoMT and increased -SMA expression were detected in PGE2- and TGF--treated HPMCs (Fig.12c).

Because PGE2 induced -SMA expression, we next sought to determine which prostaglandin receptor (EP) was responsible for PGE2-induced -SMA. HPMCS were treated with TGF- and butaprost (EP2 agonist),

sulprostone (EP3 agonist) or Cay10598 (EP4 agonist). -actin was the loading control. While butaprost and Cay10598 had no eect on -SMA expression (data not shown), sulprostone robustly induced -SMA expression (Fig.12d). We attempted to down-regulate the EP receptors (24) by siRNA transfection but were unable to sufciently reduce EP receptor expression for our analyses.

Discussion

Although the incidence of empyema continues to increase worldwide [1, 20], advances in its treatment remain relatively limited. Bacteriological analyses of the participants in the rst Multicenter Intrapleural Sepsis

Fig. 10 PGE2 and plasmin levels are increased in PAI-1 deciency after induction of S. pneumoniae mediated pleural injury. Pleural injury was induced in WT and PAI-1/ mice by intrapleural injection of S. pneumoniae and maintained for 3 days. PGE2 was extracted from the PLs of saline and S. pneumoniae-infected mice. PGE2 levels were then determined by competitive ELISA. a PGE2 was signicantly higher in S. pneumoniae infected PAI-1/ mice compared to infected

WT mice at 3 days, n = 47 mice/group. * denotes (P < 0.05). b PLs

from saline and S. pneumoniae infected WT and PAI-1/ mice were assayed for plasmin activity. S. pneumoniae infection signicantly increased plasmin activity in WT and PAI-1/ mice (p = 0.05 and

0.02, respectively) by 3 days. S. pneumoniae-infected PAI-1/ mice demonstrated signicantly higher plasmin activity than identically treated WT mice (p = 0.04)

Tucker et al. Clin Trans Med (2016) 5:17

Page 15 of 18

a

b

400

300

* *

PGE 2

(pg/mL)

200

100

0 PBS

PLN

uPA

c

d

Fig. 11 COX-2 and PGE2 are induced in human and murine PMCs by plasmin and uPA. a Serum-starved HPMCs were treated with plasmin (PLN, 7 nM) or uPA (20 nM) for 48 h. Lysates were then Western blotted for COX-2. Akt was the loading control. b Conditioned mediaof cells treated with PBS, plasmin and uPA were analyzed for PGE2.

Plasmin and uPA signicantly increased PGE2 expression in HPMCs (* denotes a p < 0.05 when compared to PBS treated controls. n = 3/

treatment). c Serum-starved MPMCs were treated with murine plasmin (7 nM, PLN) and murine uPA (20 nM) for 48 h. Lysates were then subjected to western blotting for COX-2. -actin was the loading control. d Conditioned media of cells treated with PBS, plasmin and uPA were probed for PGE2 by competitive ELISA. Plasmin and uPA signicantly increased PGE2 expression in MPMCs (*denotes a p < 0.05 when compared to PBS treated controls)

Trial (MIST1) showed that the leading causes of pleural infection are Streptococcal pathogens [1, 21, 22]. Further, the leading microbial cause of community-acquired pneumonia is now Streptococcus pneumoniae [22]. The 12 month mortality rate associated with pulmonary S.

pneumoniae infection is reported to be about 17% [22]. There has been little improvement in the mortality rate due to pleural infection in the US for the last 5 decades despite therapeutic advances including vaccination to protect against Streptococcal pneumonia [2]. These considerations suggest an imperative to better understand mechanisms that contribute to pleural injury and identify new pathways and targets amenable to intervention. These eorts have been impeded by the lack of a survivable murine model that reliably allows investigation of the pathogenesis of pleural organization and remodeling.

Our objective was to therefore develop a murine model that enables the assessment of organizing pleural injury after acute infection. We chose to develop the model in C57BL/6 mice so that investigators in the eld could use the model in genetically engineered animals that are commonly available in this strain. An alternative model of murine Streptococcal empyema was originally reported in CD1 mice by Wilkosz et al. [11]. However, the model was lethal at 48h, which precludes analyses of subsequent changes that promote pleural organization and lung restriction. Another model has been reported in rabbits intrapleurally infected with Pasteurella multocida [10]. Although this model is quite robust and amenable to intervention with brinolytic agents, a major limitation is that the organism rarely causes pleural disease in humans.

Because pleural infection most commonly occurs as a complication of pneumonia, we initially attempted to model these circumstances through nasal delivery of S. pneumoniae into the lungs and monitored the development of empyema. Intranasal administration of S. pneumoniae-induced pneumonia with severe parenchymal inammation and restrictive lung physiology, but sub-lethal doses did not reliably induce empyema, increase pleural thickening or promote pleurodesis. Further, early mortality (euthanasia required by day 2) was observed in mice with pneumonia severe enough to induce empyema. Based on histological assessments, restriction in these mice was mainly due to pneumonia rather than advanced pleural remodeling.

Because sublethal intranasal infection failed to induce pleural injury, we next administered S. pneumoniae directly into the intrapleural space. It is important to note intrapleural delivery models clinical pleural infection that may occur with direct pleural infection, as occurs after penetrating chest trauma, chest tube or thoracentesis-related infections or with post-surgical complications. This route of administration was characterized by the progressive accumulation of brinous material and increased WBCs in the pleural space by 7days after injury. Pleural organization was also apparent by CT imaging demonstrating pleural abnormalities

Tucker et al. Clin Trans Med (2016) 5:17

Page 16 of 18

Fig. 12 PGE2 induces MT of HPMCs and MPMCs. Serum-starved HPMCs were treated with PBS, PGE2 (1 M ) and TGF- (5 ng/ml) for 48 h. Cell lysates were cleared, measured and then resolved by SDS-PAGE. a -SMA was increased in TGF- and PGE2-treated HPMCs. -actin was the loading control. Image is representative of three independent experiments. b MPMCs were serum starved for 24 h and then treated with PBS, TGF- (5 ng/

ml) or PGE2 (1 M) for 48 h. Lysates were then resolved via SDS-PAGE and probed for -SMA. Akt was the loading control. Image is representative of two independent experiments. HPMCs were seeded on glass coverslips. Cells were serum starved for 24 h and then treated with PBS, PGE2 or TGF-

for 48 h. c PGE2 increased -SMA expression by HPMCs and induced changes in cell morphology indicative of MesoMT. Images were taken at 40

and are representative of 30 elds/treatment. d Serum-starved HPMCs were treated with PBS, TGF- (5 ng/ml) and sulprostone (EP3 agonist, 1 M) for 48 h. Cell lysates were cleared, measured and then resolved by SDS-PAGE. -SMA was increased in TGF- and sulprostone-treated HPMCs. -actin was the loading control. Image is representative of three independent experiments

and by signicant restrictive changes in pulmonary function by 7days. Pleural adhesion formation and thickening and areas of pleurodesis were apparent at gross inspection and by histological analyses. Although signicant decrements in lung function were observed at 3 days in infected WT mice, changes in lung volumes did not reach signicance, likely attributable to modest changes in pleural organization that were found at that time. The signicant decrements in lung volume and function demonstrated by infected PAI-1/ mice at 3 days appeared to be due to the presence of pleural eusions

rather than advanced pleural remodeling. The presence of purulent pleural eusions and increased pleural lavage neutrophilia in PAI-1/ mice compared to WT mice demonstrates that local inammation is worsened by PAI-1 deciency. Although the bacterial burden in WT and PAI-1/ was comparable at 3 days, septicemia was not excluded and represents a potential determinant for increased mortality in PAI-1 decient mice. While a range of mediators that could have contributed to increased pleural inammation in the infected PAI-1/ mice were identied, increased pleural lavage PGE2

Tucker et al. Clin Trans Med (2016) 5:17

Page 17 of 18

of these versus WT animals in particular likely contributed to that dierential response. On the other hand, increased local elaboration of plasmin likely contributed to the decreased extravascular brin seen in the PAI-1/ animals with empyema (Fig.8c), which could have impaired containment of the organisms and worsened outcomes. Lastly, it is likewise possible that local or systemic elaboration of other factors could have contributed to the increased mortality in these animals.

Antibiotic administration was required to maintain mice with empyema over 314 days after intrapleural inoculation. While ampicillin (100 mg/kg) is an alternative antibiotic for the treatment of S. pneumoniae infection, its use required subcutaneous injection every 12h. To minimize manipulation and distress in the infected mice, we chose to use the quinolone enrooxacin (15mg/ kg), which only required subcutaneous injection every 24 h. We did not attempt to induce empyema in mice pre-treated with antibiotics. Although antibiotic treatment cleared the bacterial infection by 3days, decrements in lung function and pleural rind formation occurred even after clearance of viable organisms. This situation simulates the ndings that can occur in complicated parapneumonic pleural eusions in patients that are predis-posed to loculation. As expected, a range of inammatory mediators were signicantly increased by 3 days and tended to subside by 14days after induction of empyema. Restrictive changes in lung function and pleural thickness persisted at 14days after induction of Streptococcal empyema, enabling assessment of pleural remodeling at subacute stages post-infection. While we did not carry the model forward, the animals were recovering from pleural infection at 14 days and could likely be maintained for even longer periods of time to assess the resolution of pleural injury. Studies to evaluate factors that contribute to progressive remodeling after induction of empyema over longer intervals are ongoing. Future studies will also include S. pneumoniae strains more commonly found in the clinical setting to determine if they likewise induce comparable pleural remodeling and survival.

In our previous reports, we showed that MesoMT contributed to the increased myobroblast population observed in human nonspecic pleuritis and in our carbon black/bleomycin pleural injury model [9, 17]. Further, MesoMT has been reported by others to be a key feature of pleural remodeling [23, 24]. Confocal analyses of tissue sections from S. pneumoniae-infected mice demonstrated increased -SMA expression that colocalized within the pleural and subpleural regions, demonstrating that MesoMT occurred in WT mice by 7days after pleural infection. Myobroblasts from other sources such as the lung interior or brocytes may also

be present, but the data show that mesothelial cells contribute to the process. These ndings strongly suggest that pleural mesothelial cells undergo MesoMT and contribute to pleural remodeling and rind formation that occur over 2weeks following induction of empyema.

The model is tractable for the identication of novel pathways that may condition pleural remodeling. An example of that is provided by our data showing that PGE2 is not only locally expressed during the organizational phase of resolving empyema, as expected, but that it can inuence remodeling by stimulating MesoMT and pleural rind expansion. Plasmin was detectable during this phase and was capable of inducing PGE2 elaboration by murine PMCs. Plasmin and PGE2 potently induced biomarkers of MesoMT as did TGF- in our analyses. We found that PGE2 can contribute to organization and remodeling of the visceral pleura after S. pneumoniae induced injury in part by promoting MesoMT in an EP3-specic manner. The ability of plasmin itself to induce pleural rind formation is likely limited by rapid inhibition of brinolytic activity leading to formation of an intrapleural brinous transition matrix at 7 or 14 days post infection, as demonstrated by the paucity of detectable brinolytic activity in in lung lavage.

In summary intrapleurally administered S. pneumoniae induced robust pleural remodeling. The model is also characterized by progressive matrix deposition, restrictive lung disease and durable pleural remodeling. Further, the use of antibiotics allows the study of injury progression for at least 14days and perhaps over longer periods of time. The model enables the use of genetically modied animals, as demonstrated by analyses of the eects of PAI-1 deciency in the model. We identied enhanced susceptibility to S. pneumoniae infection in PAI-1/

mice and demonstrate that pleural inammation was increased and likely due at least in part to overexpression of PGE2. Increased local elaboration of plasmin in the

PAI-1/ mice appears to have decreased extravascular brin deposition, which may have impaired containment of the infection within the pleural space. Apart from these eects, other local or systemic eects including systemic sepsis could have contributed to increased mortality in the PAI-1/ mice. In the aggregate, these ndings demonstrate that the model is reliably characterized by pleural organization after the induction of pleural infection and enables dissection of local alterations involved in the organizational phase of pleural injury that follows Streptococcal empyema.

Authors contributions

TAT, AJ, JB, BQ, SO, KBK, GF, YT, MI and AAK performed experiments presented in the manuscript. TAT and SI designed experiments presented in the manuscript. TAT, GF, AAK, and SI, prepared and approved manuscript for submission. All authors read and approved the nal manuscript.

Tucker et al. Clin Trans Med (2016) 5:17

Page 18 of 18

Acknowledgements

NIH HL115466, Seed Grant Funding from the University of Texas Health Science Center at Tyler and The Texas Lung Injury Institute.

Competing interests

The authors declare that they have no competing interests.

Received: 11 February 2016 Accepted: 3 May 2016

References

1. Lisboa T, Waterer GW, Lee YC (2011) Pleural infection: changing bacteriology and its implications. Respirology 16(4):598603. doi:http://dx.doi.org/10.1111/j.1440-1843.2011.01964.x

Web End =10.1111/j.1440-1843.2011.01964.x

2. Rosenstengel A (2012) Pleural infection-current diagnosis and management. J Thorac Dis. 4(2):186193. doi:http://dx.doi.org/10.3978/j.issn.2072-1439.2012.01.12

Web End =10.3978/j.issn.2072-1439.2012.01.12

3. Colice GL, Curtis A, Deslauriers J, Hener J, Light R, Littenberg B et al (2000) Medical and surgical treatment of parapneumonic eusions: an evidence-based guideline. Chest 118(4):11581171

4. Hener JE, Klein JS, Hampson C (2009) Interventional management of pleural infections. Chest 136(4):11481159. doi:http://dx.doi.org/10.1378/chest.08-2956

Web End =10.1378/chest.08-2956

5. Farjah F, Symons RG, Krishnadasan B, Wood DE, Flum DR (2007) Management of pleural space infections: a population-based analysis. J Thorac Cardiovasc Surg 133(2):346351. doi:http://dx.doi.org/10.1016/j.jtcvs.2006.09.038

Web End =10.1016/j.jtcvs.2006.09.038 6. Light RW (2006) Parapneumonic eusions and empyema. Proc Am Thorac Soc. 3(1):7580. doi:http://dx.doi.org/10.1513/pats.200510-113JH

Web End =10.1513/pats.200510-113JH

7. Tucker T, Idell S (2013) Plasminogen-plasmin system in the pathogenesis and treatment of lung and pleural injury. Semin Thromb Hemost 39(4):373381. doi:http://dx.doi.org/10.1055/s-0033-1334486

Web End =10.1055/s-0033-1334486

8. Idell S (2008) The pathogenesis of pleural space loculation and brosis. Curr Opin Pulm Med 14(4):310315. doi:http://dx.doi.org/10.1097/MCP.0b013e3282fd0d9b

Web End =10.1097/MCP.0b013e3282fd0d9b

9. Tucker TA, Jeers A, Alvarez A, Owens S, Koenig K, Quaid B et al (2014) Plasminogen activator inhibitor-1 deciency augments visceral mesothelial organization, intrapleural coagulation, and lung restriction in mice with carbon black/bleomycin-induced pleural injury. Am J Respir Cell Mol Biol 50(2):316327. doi:http://dx.doi.org/10.1165/rcmb.2013-0300OC

Web End =10.1165/rcmb.2013-0300OC

10. Idell S, Jun Na M, Liao H, Gazar AE, Drake W, Lane KB et al (2009) Single-chain urokinase in empyema induced by Pasturella multocida. Exp Lung Res 35(8):665681. doi:http://dx.doi.org/10.3109/01902140902833277

Web End =10.3109/01902140902833277

11. Wilkosz S, Edwards LA, Bielsa S, Hyams C, Taylor A, Davies RJ et al (2012) Characterization of a new mouse model of empyema and the mechanisms of pleural invasion by Streptococcus pneumoniae. Am J Respir Cell Mol Biol 46(2):180187. doi:http://dx.doi.org/10.1165/rcmb.2011-0182OC

Web End =10.1165/rcmb.2011-0182OC

12. Williams L, Tucker TA, Koenig K, Allen T, Rao LV, Pendurthi U et al (2012) Tissue factor pathway inhibitor attenuates the progression of malignant pleural mesothelioma in nude mice. Am J Respir Cell Mol Biol 46(2):173 179. doi:http://dx.doi.org/10.1165/rcmb.2010-0276OC

Web End =10.1165/rcmb.2010-0276OC

13. Tucker TA, Williams L, Koenig K, Kothari H, Komissarov AA, Florova G et al (2012) Lipoprotein receptor-related protein 1 regulates collagen 1 expression, proteolysis, and migration in human pleural mesothelial cells. Am J Respir Cell Mol Biol 46(2):196206. doi:http://dx.doi.org/10.1165/rcmb.2011-0071OC

Web End =10.1165/rcmb.2011-0071OC

14. Komissarov AA, Florova G, Idell S (2011) Eects of extracellular DNA on plasminogen activation and brinolysis. J Biol Chem 286(49):41949 41962. doi:http://dx.doi.org/10.1074/jbc.M111.301218

Web End =10.1074/jbc.M111.301218

15. Idell S, Zwieb C, Kumar A, Koenig KB, Johnson AR (1992) Pathways of brin turnover of human pleural mesothelial cells in vitro. Am J Respir Cell Mol Biol 7(4):414426. doi:http://dx.doi.org/10.1165/ajrcmb/7.4.414

Web End =10.1165/ajrcmb/7.4.414

16. Jeers A, Owens S, Koenig K, Quaid B, Pendurthi UR, Rao VM et al (2015) Thrombin down-regulates tissue factor pathway inhibitor expressionin a PI3 K/nuclear factor-kappaB-dependent manner in human pleural mesothelial cells. Am J Respir Cell Mol Biol 52(6):674682. doi:http://dx.doi.org/10.1165/rcmb.2014-0084OC

Web End =10.1165/ http://dx.doi.org/10.1165/rcmb.2014-0084OC

Web End =rcmb.2014-0084OC

17. Owens S, Jeers A, Boren J, Tsukasaki Y, Koenig K, Ikebe M et al (2015) Mesomesenchymal transition of pleural mesothelial cells is PI3 K and NF-kappaB dependent. Am J Physiol Lung Cell Mol Physiol 308(12):L1265 L1273. doi:http://dx.doi.org/10.1152/ajplung.00396.2014

Web End =10.1152/ajplung.00396.2014

18. Cuzzocrea S, Crisafulli C, Mazzon E, Esposito E, Muia C, Abdelrahman Met al (2006) Inhibition of glycogen synthase kinase-3beta attenuates the development of carrageenan-induced lung injury in mice. Br J Pharmacol 149(6):687702. doi:http://dx.doi.org/10.1038/sj.bjp.0706902

Web End =10.1038/sj.bjp.0706902

19. Bauman KA, Wettlaufer SH, Okunishi K, Vannella KM, Stoolman JS, Huang SK et al (2010) The antibrotic eects of plasminogen activation occur via prostaglandin E2 synthesis in humans and mice. J Clin Investig 120(6):19501960. doi:http://dx.doi.org/10.1172/JCI38369

Web End =10.1172/JCI38369

20. Burgos J, Falco V, Pahissa A (2013) The increasing incidence of empyema. Curr Opin Pulm Med 19(4):350356. doi:http://dx.doi.org/10.1097/MCP.0b013e3283606ab5

Web End =10.1097/MCP.0b013e3283606ab5

21. Maskell NA, Davies CW, Nunn AJ, Hedley EL, Gleeson FV, Miller R et al (2005) U.K. controlled trial of intrapleural streptokinase for pleural infection. N Eng J Med 352(9):865874

22. Maskell NA, Batt S, Hedley EL, Davies CW, Gillespie SH, Davies RJ (2006) The bacteriology of pleural infection by genetic and standard methods and its mortality signicance. Am J Respir Crit Care Med 174(7):817823. doi:http://dx.doi.org/10.1164/rccm.200601-074OC

Web End =10.1164/rccm.200601-074OC

23. Decologne N, Wettstein G, Kolb M, Margetts P, Garrido C, Camus P et al (2010) Bleomycin induces pleural and subpleural brosis in the presence of carbon particles. Eur Respir J 35(1):176185. doi:http://dx.doi.org/10.1183/09031936.00181808

Web End =10.1183/09031936.00181808

24. Decologne N, Kolb M, Margetts PJ, Menetrier F, Artur Y, Garrido C et al (2007) TGF-beta1 induces progressive pleural scarring and subpleural brosis. J Immunol 179(9):60436051