Targeting the PAI-1 Mechanism with a Small Peptide Increases the Efficacy of Alteplase in a Rabbit Model of Chronic Empyema

The incidence of empyema is increasing and associated with a mortality rate of 20% in patients older than 65 years. Since 30% of patients with advanced empyema have contraindications to surgical treatment, novel, low-dose, pharmacological treatments are needed. A Streptococcus pneumoniae-induced rabbit model of chronic empyema recapitulates the progression, loculation, fibrotic repair, and pleural thickening of human disease. Treatment with single chain (sc) urokinase (scuPA) or tissue type (sctPA) plasminogen activators in doses 1.0–4.0 mg/kg were only partially effective in this model. Docking Site Peptide (DSP; 8.0 mg/kg), which decreased the dose of sctPA for successful fibrinolytic therapy in acute empyema model did not improve efficacy in combination with 2.0 mg/kg scuPA or sctPA. However, a two-fold increase in either sctPA or DSP (4.0 and 8.0 mg/kg or 2.0 and 16.0 mg/kg sctPA and DSP, respectively) resulted in 100% effective outcome. Thus, DSP-based Plasminogen Activator Inhibitor 1-Targeted Fibrinolytic Therapy (PAI-1-TFT) of chronic infectious pleural injury in rabbits increases the efficacy of alteplase rendering ineffective doses of sctPA effective. PAI-1-TFT represents a novel, well-tolerated treatment of empyema that is amenable to clinical introduction. The chronic empyema model recapitulates increased resistance of advanced human empyema to fibrinolytic therapy, thus allowing for studies of muti-injection treatments.


Introduction
For several decades, the incidence of empyema [1] steadily rose, together with mortality [2][3][4][5][6][7][8], in the United States, with an annual cost of care of up to five hundred million dollars [9]. According to the American Thoracic Society, empyema advances through three sequential stages-from early exudative (I) to fibrinopurulent (II) and finally to organization (III), which features adhesions sometimes leading to lung entrapment and a thick pleural rind. While the therapeutic goal, re-expansion of the affected lung, is achievable with thoracentesis in stage I, more advanced empyema (stage II) requires video-assisted thoracoscopic surgery (VATS) or intrapleural fibrinolytic therapy (IPFT) and stage III often requires thoracotomy [8,[10][11][12][13][14][15]. Patients that present with extensive loculation, lung abscess, and pleural thickening are more likely to suffer failed IPFT [16][17][18][19][20][21] and require surgical intervention [19,22]. While thoracoscopic surgery is a widely accepted primary therapy for empyema [15], up to 30% of patients are poor candidates for either surgery or, alternatively, for conventional IPFT. Moreover, the cohort of patients who are treated only with simple drainage were a decade older and had higher mortality than those who are good candidates for the thoracoscopic surgery [14,23,24]. Pharmacological treatment  (24 h). Infectious pleural injury was initiated by intrapleural injection of S. pneumoniae at time zero. Progression of the pleural fibrosis was monitored by daily ultrasonography [41,42]. The acute phase (yellow) approaches its peak by 3 d. The chronic phase, which involves increased intrapleural organization (red), follows the acute phase. Both standard Intrapleural Fibrinolytic Therapy (IPFT) with sctPA or scuPA or PAI-1-TFT (green) were given at 7 d after initiation of empyema. Excess pleural fluid (up to 10 mL) was drained as needed at 8 h after treatment. Outcomes were assessed at 8 d (24 h after treatment) using ultrasonography and postmortem visualization (documented by photography). Samples of lung tissue were collected and stained by hematoxylin and eosin (H&E). A Gross Lung Injury Score (GLIS) [40][41][42], which ranges from 0 to 50 units, where a clear pleural space is 0; too numerous to count (TNTC) fibrin formations corresponds with 50, and GLIS ≤ 10 indicates successful treatment, was determined. GLIS is the sum of the number of intrapleural fibrin strands (1 unit each) and large nets and aggregates (5 units each). (b) (A) Chest ultrasonography of a S. pneumoniae-induced chronic empyema model at 7 d. The yellow arrows indicate intrapleural fibrin deposition; L = lung, H = heart; (B) postmortem visual evaluation with extensive adhesions, fibrinous coating of the lung, and multiloculated right hemithorax; (C) histologic assessment shows a pleural surface coated with fibrin, inflammation, and subpleural pneumonitis with increased organization. Pleural thickening (black arrows) surface to basement membrane (1 mm scale); (D) histologic assessment of the edge of a normal, uninjured lung with pleural thickening (black arrows) and a scale bar 1 mm. The visceral pleural surface is oriented at the bottom left portion of each panel. (c) Changes in the pleural fluid levels of PAI-1 and biomarkers of inflammation during the transition from acute to chronic empyema in rabbits. Samples of pleural fluid were withdrawn at 3 d (baseline acute, early stage) and 7 d (baseline chronic, advanced-stage) and levels of total (A) and active (B) PAI-1, TGF-β (C), TNF-α (D), IL-6 (E), IL-8 (F), and IL-1β were determined as previously described [41]. Statistical significance between these two sample groups was determined using an unpaired, 2-tailed Kolmogorov-Smirnov test. Statistical significance: **** and ns (not significant) denote p < 0.0001 and > 0.05, respectively. ELISA (R&D Systems, IN; Molecular Innovations, MI; and Ray Biotech, GA) were used to determine levels of proteins in pleural fluids.

Animal Protocols
Animal procedures and techniques were approved by the Institutional Animal Care and Use Committee at The University of Texas Health Science Center at Tyler (IACUC protocols 616, 672). New Zealand White rabbits (2.9-3.6 kg; average age 18 weeks) from Charles River Laboratories (Wilmington, MA, USA) were used for the model of chronic empyema. Female, pathogen-free animals (n = 46) were required for these experiments. An infectious pleural injury in rabbits was induced as described elsewhere [41,42]. Briefly, 1-5 × 10 8 cfu of S. pneumoniae (D39 strain, National Collection of Type Cultures, Salisbury UK) in 3 mL of 0.5% brain-heart infusion agar (BD 238400, BD Diagnostic Systems, Hunt Valley, MD, USA) was injected into the right pleural space of an animal. Clavomox (10 mg/kg, subcutaneous, daily for 1-3 days as clinically indicated by Attending Veterinarian) (10000485, Zoetis, Parsippany, NJ, USA) was started at 28-30 h post-infection. Pleural injury and accumulation of loculation and pleural fluid was monitored by daily ultrasonography. PAI-1-TFT was administered via an 18-gauge catheter which was flushed with phosphate-buffered saline (0.5 mL). Samples of pleural fluids (≥0.5 mL) were collected prior to (baseline) and at 8 and 24 h after treatment [39,40,42,45]. Anesthesia, postoperative pain medication, and animal care were provided as reported previously [39,40,42,45]. Animals were monitored for signs of distress, pain, and worsening clinical status. In absence of these, rabbits were maintained for six days, treated on the seventh day, and euthanized on the eighth day so that the pleural space could be imaged, and fluid and tissue samples could be collected. Euthanasia was accomplished using intravenous injection of 1 mL of commercial euthanasia solution (sodium pentobarbital 390 mg/mL and phenytoin 50 mg/mL) followed by exsanguination.

Ultrasonography
Development of chronic empyema was monitored via B-mode ultrasonography of the chest [39,46] using a Logiq e system (GE Healthcare, Milwaukee, WI, USA) equipped with version R5.2.x software and a multifrequency transducer model 12L-RS at a working frequency of 10 MHz, as previously reported [42].

Metrics of Treatment Efficacy and Pleural Injury
Gross Lung Injury Score (GLIS) was determined for each animal at 24 h after treatment (8th day). Post-euthanasia, the rabbit pleural cavity was opened, and pathological structures noted. GLIS varied from 0 (clear pleural space) to 50 (TNTC, too numerous to count), and equaled to a sum of fibrin strands (score 1 per each discrete strand), webs, sheets, and large (>5 mm) aggregates (score 5) as previously described [39,41,45]. Multiple visceral-parietal interconnected fibrin sheets and/or webs corresponded to GLIS = 50. PAI-1-TFT was considered successful if GLIS ≤ 10. Minimal effective dose (MED) is defined as the minimal dose of a plasminogen activator, alone or with PAI-1-targeting adjunct, which results in GLIS ≤ 10 for every animal in a group (n = 5-6). Morphometry was used to determine pleural thickening as reported elsewhere [42].

Histology
Tissue samples from rabbit lungs were paraffinized and sectioned to be stained. Hematoxylin and Eosin (H&E) (StatLab, Columbia, MD, USA) and Masson Trichrome (Thermo Fisher, Waltham, MA, USA) staining techniques were used to visualize pleural thickening and presence of collagen. Color bright field imaging at 4× were performed using Cytation 5 (Biotek, Winooski, VT, USA).

Data Analysis and Statistics
Levels of statistical significance for non-pairing groups greater than two were determined using a two-tailed Kruskal-Wallis test with Dunn's multiple comparison test. The level of statistical between two nonpaired groups were determined via two-tailed Kolmogorov-Smirnov for cumulative distribution and two-tailed Mann-Whitney test for cumulative ranks. A paired t-test was used to determine the statistical significance of paired data. Data analysis was performed using GraphPad Prism 9.3.1 as previously described [39][40][41][42].

Chronic Infectious Pleural Injury in Rabbits Recapitulates Advanced-Stage Empyema in Humans
A model of chronic empyema ( Figure 1) includes the progression of infectious pleural injury beyond the 4 d acute [41] stage up to 21 days after induction of injury, and features an increase in the severity of pleural fibrosis and pleural thickening [42]. An 8 d model (7 d development followed by 24 h treatment) of chronic empyema (Figure 1, panel (a)) was chosen to evaluate the efficacy of bolus injections of either scuPA or sctPA and determining the MED for each fibrinolysin. This model of advanced-stage, chronic empyema features robust intrapleural fibrin deposition, detected by ultrasonography ( Figure 1, panel (b), A), and documented by photography during necropsy ( Figure 1, panel (b), B). Notably, the severity of the intrapleural fibrin and collagen deposition prevented pleural fluid drainage and progressed further than in models of chemically induced pleural injury and acute, early-stage empyema [41]. Pleural thickening in this rabbit model of advanced-stage empyema was also increased when compared to acute empyema in rabbits and chemically induced pleural injury [41] and to naïve rabbit lungs ( Figure 1, panel (b), C and D). The timeline of advanced-stage infectious plural injury development and treatment are shown in Figure 1, panel (a). Changes in baseline levels of PAI-1 (total and active) and biomarkers of inflammation (TGF-β, TNF-α, IL-6, and IL-8) in pleural fluids of animals with advancedstage (7 d) pleural injury, when compared to the acute stage (3 d) are shown as box plots in Figure 1 (panel (c), A-F, respectively). Notably, while levels of total PAI-1, TNF-α, IL-6, IL-8 ( Figure 1, panel (c)), and IL-1 β (below level of detection), at baseline in chronic empyema were lower (p < 0.05) than those in the acute stage, the level of active PAI-1 did not change significantly, and the level of TGF-β was elevated (p < 0.05) when compared to baseline in acute empyema in rabbits [41] (Figure 1, panel (c)). During the transition between acute (3 d) and chronic (7 d) empyema in rabbits (Figure 1, panel (a)), the levels of inflammatory markers TNF-α, IL-6, IL-8, and IL-1 β in the pleural fluid at baseline decreased by 2.5, 57, 2.1, and more than 100-fold, respectively. In contrast, the level of TGF-β increased by 1.5-fold, likely promoting further development of intrapleural fibrosis, and increasing pleural thickening (Figure 1, panel (b), C). While the level of total PAI-1 decreased by 5.6-fold, the level of active PAI-1 present in the pleural space did not change (Table A1). Thus, levels of both active and total PAI-1, our molecular target, in chronic, advanced-stage empyema in rabbits remains up to two orders of magnitude higher than those in uninjured animals. Levels of TNF-α, IL-6, IL-8 in chronic empyema model were also markedly higher than those observed in uninjured animals [49,50]. Moreover, levels of PAI-1, TGF-β, TNF-α, IL-6, IL-8 observed in pleural fluids of human patients with empyema and with parapneumonic effusions were also higher than those in the transudative pleural fluids and were comparable to the levels observed in the rabbit empyema model (Table A1). At the time of treatment (7 d; Figure 1, panel (a)), both fibrinolytic and plasminogen-activating (PA) activities in pleural fluids collected from animals were suppressed by overexpressed PAI-1 ( Figure A1). However, supplementation of the pleural fluid with exogenous tPA resulted in fibrinolytic activity due to activation of accumulated endogenous plasminogen ( Figure A1). There was no statistical difference in red (RBC) or white (WBC) blood cell counts in the pleural fluids from animals with acute and chronic phases of S. pneumoniae-induced pleural injury. Nevertheless, there was a statistically significant decrease (p < 0.05) in the neutrophil count, with a decrease in neutrophil-lymphocyte ratio noted in pleural fluids of chronic phase of empyema in rabbits. These changes in the degree of pleural organization and inflammation may affect the processing of fibrinolysins during IPFT of advanced-stage empyema and result in an increase in the time needed for effective fibrinolysis.

The Rabbit Model of Chronic Empyema Recapitulates the Decrease in the Efficacy of Fibrinolytic Therapy Observed in Humans with Advanced-Stage Empyema
Since the level of the molecular target, PAI-1, in chronic (7 d) empyema in rabbits was significantly lower than that in the acute stage ( Figure 1, panel (c), A), a dose at half of the MED determined in the acute model [41,42] was chosen as the starting dose for treatment of chronic empyema in this rabbit model. S. pneumoniae was injected into the pleural space on 0 d (Figure 1, panel (a)) and the progression of the infectious pleural injury was monitored using daily ultrasonography ( Figure 1, panel (a)). Intrapleural bolus injection of human sctPA or scuPA was administered on 7 d. Samples of pleural fluid were drawn prior to (baseline) and at 8 h and 24 h after treatment. The efficacy of treatment of chronic empyema was monitored by ultrasonography with postmortem visualization (documented by photography) and histological analysis of the collected lung tissue ( Figure 2, panel (b)). Gross Lung Injury Scores (GLIS) [40][41][42] were used to determine the efficacy of the treatment (Figure 2, panel (a), A). Successful treatment outcome had GLIS ≤ 10 [40][41][42]. A starting dose of 1.0 mg/kg sctPA (n = 3) was not efficacious as expected for the MED and the dose was escalated by two-fold (bolus injection, 2.0 mg/kg) of sctPA (n = 6) or scuPA (n = 5)). However, neither plasminogen activator demonstrated the efficacy required for the MED (Figure 2, panel (a)). In order to identify the MED for a bolus injection, the dose was doubled again to 4.0 mg/kg, which is comparable to doses of Alteplase [37] and Urokinase [51] used in clinical practice. However, neither sctPA (n = 5), nor scuPA (n = 5) resulted in a treatment outcome of MED (GLIS ≤ 10 for every animal in a group, n = 5-6) ( panel (a)). There was no statistically significant difference (p > 0.05) in RBC or WBC counts in the pleural fluids at 24 h after treatments ( Figure A2). Levels of PAI-1 (total and active) and inflammatory biomarkers in pleural fluids collected at 24 h after treatment are shown in Figure A3. These higher doses of treatment with a bolus injection of sctPA or scuPA approached the highest doses (5.0-20.0 mg per injection) of Alteplase and Urokinase, which are currently used clinically [37,38,[51][52][53]. However, treatment of chronic, advanced-stage empyema with 12.0-13.2 mg of either fibrinolysin did not reach the efficacy of MED. Thus, this model of advanced-stage empyema recapitulates failure of IPFT with an increase in severity of empyema observed in humans, and calls for testing of the multiple injections of plasminogen activators, as is carried out in clinical practice [37,38,[51][52][53]. Notably, similar to the acute empyema model [42], successful treatment with sctPA or scuPA did not result in a decrease in the pleural thickening (

DSP-Based PAI-1-Targeted Fibrinolytic Therapy Increases the Efficacy of sctPA in a Rabbit Model of Chronic Empyema
Targeting the PAI-1 mechanism in chronic empyema allows for further validation of the approach and testing of the concept that PAI-1-TFT improves therapeutic outcomes in empyema. The efficacy of sctPA in an acute model empyema was increased by up to eight-fold using PAI-1-TFT with 8.0 mg/kg DSP [41]. Thus, 8.0 mg/kg DSP in combination with ineffective doses of 2.0 mg/kg sctPA or scuPA was tested to evaluate the effect of PAI-1-TFT on single bolus injection treatment of chronic empyema in rabbits ( . Thus, targeting the PAI-1 mechanism with DSP did not improve outcomes of scuPA treatment but rendered ineffective doses of sctPA effective. Previously, we demonstrated that PAI-1-TFT was effective for treatment of chemically induced pleural injury and in a model of acute (4 d), early-stage empyema [40,41,43] (Table 1). While DSP mediated PAI-1-TFT was effective with both sctPA and scuPA in treatment of chemically induced pleural injury, it was effective only with sctPA in acute and chronic stages of the infectious model (Table 1). However, MED for sctPA or scuPA alone in the chronic empyema model exceeds the doses currently used in clinical practice, thus bringing no advantage to further dose escalation experiments. There was no statistically significant difference (p > 0.05) in RBC or WBC counts in the pleural fluids at 24 h after treatment with DSP-based PAI-1-TFT ( Figure A2).      [40,41,43], Acute [41], and Chronic Infectious Pleural Injury in Rabbits. While levels of PAI-1 (total and active) and inflammatory biomarkers in pleural fluids collected at 24 h after treatment with DSP mediated PAI-1-TFT were not statistically significantly different (p > 0.05), there was a trend towards a decrease in each biomarker successful treatment with DSP combined with sctPA ( Figure A4). Thus, targeting inflammation simultaneously with PAI-1 targeting may increase the efficacy of DSP mediated PAI-1-TFT in the chronic empyema model.  [40][41][42][43]46]. Since both plasminogen-activating (Figure 4, panel (a), E) and fibrinolytic activities were suppressed at 24 h after treatment in pleural fluids of animals with both successful (GLIS ≤ 10) and unsuccessful (GLIS > 10) treatment, the TEF for chronic empyema was between 8 and 24 h. PAI-1 and fibrin were colocalized on the surface of the lung in chronic infectious pleural injury modes (Figure 4; panel (b)); this association could protect PAI-1 from immediate inactivation by an excess of plasminogen activator at IPFT. Thus, modifications of fibrin structure and a molecular target, PAI-1, entrapped in fibrin in chronic empyema may adversely affect the rate of intrapleural fibrinolysis resulting in failure of IPFT. Since further single bolus dose escalation (Figure 2, panel (a), A) in the model of advanced stage empyema could potentially increase bleeding, a strategy of multiple injection via a chest tube, similar to the approach used in clinical practice, may prove to be a more rewarding approach to be tested in future studies.  D2); (E) High levels of active plasminogen activator in samples of pleural fluid collected at 8 h after treatment with sctPA (black bar) or scuPA (gray bar) together with 6-8-fold higher levels of fibrinolysin antigens (hatched bars). At 24 h after treatment, plasminogen activating activity was completely suppressed (below level of detection), although antigen is still present. (b) PAI-1 is detectable in fibrin organizing in advanced-stage (7 d) empyema in rabbits. Fixed rabbit lungs with chronic empyema were paraffinized, sectioned, and attached to adhesive slides. Double-labeled immunofluorescence staining were performed, probing for nuclear stain (DAPI, Hoechst 33342), PAI-1 (GFP, AF 488), and Fibrinogen (CY5, AF 647). Imaging was performed at 4× (A-D) and 20× (E) magnification using Cytation 5. (c). Visualization of PAI-1 expressing activated mesothelial cells embedded in a fibrin scaffold. Chronic empyema pleural fluids collected from diseased rabbits were three dimensionally cultured in a sterile 8-well EZ slide (Millipore Sigma, Inc., Burlington, MA, USA) for up to 7 days. Double-labeled immunofluorescence staining was performed on the cells, as described. Z-stack imaging was performed at 40× (A-D) magnification using Cytation 5.

Discussion
PAI-1, which is elevated by up to three orders of magnitude in pleural fluid in empyema [32][33][34], was recently identified as a biomarker for septation severity and poor patient outcomes in empyema [44,61]. The goals of the present study were validation of a

Discussion
PAI-1, which is elevated by up to three orders of magnitude in pleural fluid in empyema [32][33][34], was recently identified as a biomarker for septation severity and poor patient outcomes in empyema [44,61]. The goals of the present study were validation of a rabbit model of advanced-stage empyema and further testing of our central hypothesis, that PAI-1 is a useful molecular target in infectious pleural injury treatment. At time point 0, both fibrinolytic and plasminogen activating activity in pleural fluids from both humans [38,62] and rabbits [40,41,43,62] (Table A1) were suppressed by highly overexpressed PAI-1 [41,42], Figure 1c panels A, B as well as Figures A3 and A4 (panels A and  B). However, plasminogen accumulated in the pleural fluid, when activated to plasmin by tPA or uPA, produced fibrinolytic activity ( Figure A1). S. pneumoniae pleural injury in rabbits [41,42] recapitulates key features of empyema in humans, including staging and timing, progressing from an early acute to more severe stage, which is accompanied by increasing pleural fibrosis, loculation, and pleural thickening. Validation of the S. pneumoniae-induced rabbit model of chronic infectious pleural injury is vital for further translation of preclinical observations to clinical trials in human patients. Demonstrating that the model mirrors the pathogenesis of empyema that was observed in clinical settings establishes the face validity of the model [63]. The model recapitulates advanced-stage empyema in humans, levels of intrapleural biomarkers of inflammation and PAI-1, showing increasingly severe pleural injury and pleural thickening in combination with resistance to bolus injection fibrinolytic therapy with high doses of sctPA and scuPA. Thus, the predictive validity of the model was established by treating animals with human plasminogen activators in doses comparable to those currently used in clinical settings [37,38,[51][52][53]. Treatment with sctPA (1.0-4.0 mg/kg) was successful (GLIS ≤ 10) in approximately 50% of animals, and the efficacy of the highest dose of sctPA (4.0 mg/kg) was statistically different (p < 0.05) when compared to the vehicle control. In contrast to sctPA, scuPA (2.0-4.0 mg/kg) was markedly less effective. These results clearly support testing a multi-dose (multiple injections) approach to treatment with fibrinolysins to further increase the efficacy of IPFT and identify a MIMED (multiple injection minimal effective dose). Multiple intrapleural injections of Alteplase (sctPA) [37,52] and Urokinase [38,51,53] are currently used to treat empyema in humans.
DSP is a short, negatively charged peptide (EEIIMD). Being part of the primed side of the reactive center loop of PAI-1, it presumably interferes with exosite interactions [64,65] between the 37-loop of the enzymes and PAI-1 [66,67]. Previously, we demonstrated that PAI-1-TFT with DSP or PAI-1 neutralizing mAbs increased the efficacy of sctPA up to eight-fold in two rabbit models of pleural injury (chemically induced and acute infectious) [40,41,43]. Here, we demonstrated that targeting the PAI-1 mechanism with DSP in a model of chronic, infectious empyema increases the efficacy of the treatment and brings the MED of sctPA with 16.0-8.0 mg/kg DSP to 2.0-4.0 mg/kg (Figure 3; panel (a) A) and predicts a better outcome if doses of both sctPA and DSP are elevated (to 4.0 and 16.0 mg/kg, respectively). In contrast to sctPA, DSP did not improve outcomes of fibrinolytic therapy with scuPA in either acute [41] or chronic (Figure 3, panel (a) A) empyema models. The higher affinity of tPA to PAI-1 compared to uPA, which depends on the protonation state of a histidine residue [68], may contribute to the striking difference between outcomes of DSP-mediated PAI-1-TFT with sctPA and scuPA in empyema models. The results of PAI-1-TFT with DSP and sctPA clearly demonstrate that the PAI-1-mechanism is a validated molecular target in both acute- [41], and chronic, advanced-stage ( Figure 3) empyema modeled in rabbits. Previously, an increase in the MED of sctPA and an increase in the dose of DSP was required for 100% effective therapy (GLIS ≤ 10), which correlated with an increase in the level of PAI-1 in acute empyema, when compared to chemically induced pleural injury [41]. However, the decrease in the level of the molecular target in pleural fluid observed in chronic empyema, when compared to the acute stage (Figure 1, panel (c), A), did not result in a decrease in the MED of sctPA. The efficacy of both fibrinolysins (sctPA and scuPA) alone (Figure 2, panel (a), A) or in combination with DSP (Figure 3; panel (a), A) in the treatment of chronic empyema was markedly lower than in acute empyema [41]. Moreover, while the dose escalation of up to 4.0 mg/kg resulted in a statistically significant difference (p < 0.05) between sctPA-treated and vehicle control groups, the treatment outcome never approached the efficacy required by the definition of MED with any fibrinolysin. sctPA or scuPA alone or DSP-based PAI-1-TFT, successfully clears intrapleural fibrin deposition, it does not significantly impact the pleural thickening either in acute, early-stage [41,42] or chronic, advanced-stage empyema in rabbits. Likewise, treatments in the chronic empyema model caused no effect on pleural thickening when compared to vehicle controls at 24 h ( Figure 2 panel (a), B; Figure 3 panel (a), B). Notably, this rabbit model of chronic empyema informs on possible limitations of PAI-1 as a biomarker of severity of intrapleural septation and poor outcomes [44]. While intrapleural levels of PAI-1 in a model of acute, early-stage empyema were higher than those in chronic, advanced-stage empyema (Figure 1, panel  (c), A), the efficacy of both IPFT and PAI-1-TFT was markedly lower in treatment of the latter (Figure 2, panel (a); Figure 3; panel (a)) [41]. These results recapitulate advancedstage empyema in humans [17,18,20], where IPFT fails more frequently due to slower fibrinolysis [20,21,30,69,70].
Thus, the rabbit model of S. pneumoniae-induced pleural injury recapitulates the evolution of untreated empyema from an acute to a more chronic stage with severe fibrosis, multiple loculation, pleural thickening, and a notable decrease in the efficacy of both IPFT and PAI-1-TFT. Compensating for slow fibrinolysis by increasing the dose of sctPA or scuPA (Figure 2, panel (a)), the frequency of multiple injections may result in only a slight increase in the efficacy, which may not compensate for the increased risk of bleeding complications. Indeed, treating patients with advanced-stage empyema using an "extended MIST2 protocol" (Alteplase 10 mg/DNase 5 mg, twice daily, for >6 doses) resulted in a four-fold increase in bleeding complications (10% compared to 2.5% using the standard protocol) [71]. Thus, the validity of our model is supported by its similarity to clinical presentation of advanced-stage empyema and responses to fibrinolytic therapy with humantailored drugs, recapitulating findings observed in a clinical setting.
Two factors, the intrapleural half-life of active fibrinolysin and time needed for effective fibrinolysis, dramatically contribute to the outcome of pharmacological treatment of empyema [41,43,72]. Interestingly, the same combination of high intrapleural plasminogenactivating activity and progressing fibrinolysis, which is detected by ultrasonography at 8 h after treatment with a high dose of plasminogen activator, could be present at 24 h in both effective and ineffective outcomes (Figure 4, panel (a)). The transition of pleural injury from an acute to a chronic stage results in a decrease in PAI-1 and maturation of the intrapleural fibrin with the development of scarring and collagen deposition [73], together with incorporation of nucleic acids into fibrin structure [48], adversely affects the rate of intrapleural fibrinolysis. However, PAI-1 incorporated in fibrin may be protected from immediate inactivation by injected plasminogen activator ( Figure 4, panel (b)), and it may play a critical role in the failure of IPFT in advanced-stage empyema, when more mature organization is observed. Thus, PAI-1-TFT affects the PAI-1 mechanism, protecting fibrinolysin from inhibition by any form of active PAI-1, increasing intrapleural half-life of the fibrinolysin, and supporting slow fibrinolysis. Therefore, PAI-1-TFT could be a reasonable alternative to dose escalation, compensating for the combination of slow fibrinolysis, rapid clearance of fibrinolysin from the pleural space, and constant replenishment of PAI-1 due to local overexpression. Table 1 demonstrates that (i) bolus injection MED for sctPA or scuPA alone in chronic, advanced-stage empyema in rabbits was higher than 4.0 mg/kg (average 13.2 mg which is comparable to doses using in clinical practice); (ii) DSP-mediated PAI-1-TFT increased the efficacy of sctPA and may convert ineffective doses of sctPA (2.0 and 4.0 mg/kg) to a MED (100% successful outcome in group of n = 5-6). Notably, the maximal decrease in MED by targeting the PAI-1 mechanism with DSP was eight-fold in both chemically induced and acute infectious models of pleural injury [41]. Moreover, the same eightfold increase in the efficacy of scuPA was observed using mAbs-mediated PAI-1-TFT in chemically induced pleural injury [40] (Table 1). If eight-fold is the limit of efficacy increase and MED of sctPA for PAI-1-TFT with 16.0 mg/kg DSP is 2.0 mg/kg (Figure 3, panel (a) A; Table 1), the expected MED for sctPA alone in chronic empyema in rabbits could be as high as 16.0 mg/kg or more. This dose (~50 mg for bolus injection) is 5-fold higher than the dosage of Alteplase and Urokinase currently used in clinical settings and could induce significant intrapleural bleeding or even enter the bloodstream, activating the fibrinolytic system in the circulation. Novel, low-dose PAI-1-TFT may minimize the risk of bleeding complications and other adverse effects of high doses of plasminogen activator and allows treatment to be adapted to the population of patients with comorbidities and contraindications to conventional IPFT and surgery. These findings provide further support to our concept that novel strategies of targeting the PAI-1 mechanism could result in more effective clearance of pleural septations and improve therapeutic outcomes with lower doses of fibrinolysin [40,41,43].

Conclusions
Preclinical development of effective low-dose treatments for patients with chronic empyema, who decline or are not candidates for surgical intervention and are unable to receive conventional high-dose IPFT, could result in a paradigm shift in empyema management. The major conclusions from this work are:

1.
Face and predictive validation of a novel model of chronic, advanced-stage empyema that closely recapitulates key physiological indicators seen in empyema in humans; 2.
PAI-1 is a valid molecular target in S. pneumoniae-induced empyema modeled in rabbits, which closely recapitulates the key characteristics of this disease in humans; 3.
Multiple injections of fibrinolysins could be tested in this model in order to develop and validate novel approaches to treatment of chronic empyema and expand treatment to patients who are not currently candidates for IPFT or surgery; 4.
A new form of the molecular target, active PAI-1 that is incorporated into a fibrin mesh could contribute to slow fibrinolysis and failure of IPFT in treatment of advancedstage empyema; 5.
Preclinical development and testing of novel, low-dose PAI-1-TFT could result in innovative approaches to treating patients with advanced-stage, organizing pleural injury that fails drainage and is untreatable with surgery or high dose IPFT, as well as creating a foundation for translation to clinical trials.
Any improvement in non-surgical drainage by decreasing the dose of plasminogen activator, increasing the rate of intrapleural fibrinolysis, or increasing the half-life of plasminogen activating and fibrinolytic activities could positively affect the survival of patients with contraindications to or elevated risk from currently available interventions for chronic empyema. Even small improvements in the current treatment of advanced-stage empyema could result in a rise in efficacy and decrease mortality in this patient cohort.

Conflicts of Interest:
Drs. Komissarov and Florova are supported by NIH and serve as co-investigators on research involving intellectual property licensed to Lung Therapeutics, Inc. (LTI) and have conflictof-interest management plans at The University of Texas Health Science Center at Tyler (UTHSCT). Dr. Idell is likewise supported by NIH, founded Lung Therapeutics, Inc., and has equity in the company, which is commercializing single chain urokinase for empyema, serves as a paid consultant as needed for the company and has a conflict-of-interest management plan from the University of Texas at Tyler accordingly. Mr. Messrs Girard and Mr. Sarva served as research associates on research involving intellectual property licensed to LTI and likewise had conflict-of-interest management plans at UTHSCT. The rest of the authorship has no conflicts of interest to disclose.
Appendix A Table A1. Levels of PAI-1, TGF-β, TNF-α, IL-6, and IL-8 in pleural fluid of rabbit empyema model and humans with complicated parapneumonic effusion/empyema, parapneumonic, and transudative pleural effusions.  , and measurements of standards with known concentrations as previously described [39,47]. Fibrinolytic activity (FA) was measured using 96-well plates with FITC-fibrin film as previously described [40,42,43]. Fibrinolytic activity in pleural fluids was calculated from the slopes of an increase in the fluorescence emission at 512 nm (excitation at 490 nm) over time as the FITC-fibrin degrades. Fibrinolytic potential was determined as the difference between fibrinolytic activity in samples of pleural fluid at baseline measured with (FA/tPA) and without (FA) supplementation with 10 nM sctPA [40]. Statistically significant differences between outcomes were determined using a two-tailed, paired t-test; **** denotes p < 0.0001. , and measurements of standards with known concentrations as previously described [39,47]. Fibrinolytic activity (FA) was measured using 96-well plates with FITCfibrin film as previously described [40,42,43]. Fibrinolytic activity in pleural fluids was calculated from the slopes of an increase in the fluorescence emission at 512 nm (excitation at 490 nm) over time as the FITC-fibrin degrades. Fibrinolytic potential was determined as the difference between fibrinolytic activity in samples of pleural fluid at baseline measured with (FA/tPA) and without (FA) supplementation with 10 nM sctPA [40]. Statistically significant differences between outcomes were determined using a two-tailed, paired t-test; **** denotes p < 0.0001.  , IL-6 (E) and IL-8 (F) were determined as previously described [41]. Statistically significant differences between levels of biomarkers were determined using a two- Figure A3. Levels of PAI-1 and Inflammatory biomarkers in pleural fluid of rabbits at baseline and 24 h after IPFT with sctPA or scuPA. Samples of pleural fluid were collected from animals with advanced-stage infectious pleural injury (from left to right) at the baseline (n = [22][23][24][25][26], and at 24 h after treatment with vehicle control (n = 6); 1.0 mg/kg sctPA (n = 3); 2.0 mg/kg sctPA (n = 6), 2.0 mg/kg scuPA (n = 5); 4.0 mg/kg sctPA (n = 5), 4.0 mg/kg scuPA (n = 5). Levels of total (A) and active (B) PAI-1, TGF-β (C), TNF-α (D), IL-6 (E) and IL-8 (F) were determined as previously described [41]. Statistically significant differences between levels of biomarkers were determined using a two-tailed Kruskal-Wallis test with Dunn's multiple comparison test; * and **** denote p < 0.05 and 0.001, respectively. ELISA (R&D Systems, IN; Molecular Innovations, MI; and Ray Biotech, GA) were used to determine levels of proteins in pleural fluids.
tailed Kruskal-Wallis test with Dunn's multiple comparison test; * and **** denote p < 0.05 and 0.001, respectively. ELISA (R&D Systems, IN; Molecular Innovations, MI; and Ray Biotech, GA) were used to determine levels of proteins in pleural fluids. (E) and IL-8 (F) were determined as previously described [41]. Statistically significant differences between levels of biomarkers were determined using a two-tailed Kruskal-Wallis test with Dunn's multiple comparison test; ** denotes p < 0.01. ELISA (R&D Systems, IN; Molecular Innovations, MI; and Ray Biotech, GA) were used to determine levels of proteins in pleural fluids.