Pancreatitis-associated Protein Protects the Lung from Leukocyte-induced Injury 

The techniques of preparing and perfusing isolated rabbit lungs have been previously described in detail. 18,19After approval by the local board for animal protection, female chinchilla rabbits (Orticolagus caniculus ) weighing 2,100 ± 196 g (mean ± SD) were anesthetized with ketamine (50 mg/kg) and xylazine (4 mg/kg) and anticoagulated with heparin–sodium 1,000 U/kg injected in the ear vein. After placement of a tracheostomy tube, the rabbits were mechanically ventilated with room air. The thorax was opened via  the diaphragm, and after a median sternotomy, a catheter was inserted into the pulmonary artery. The lung organ preparation was isolated and suspended from a weight transducer in a temperature-controlled (37°C) and humidified chamber. After the cannulation procedure, the lungs were perfused with 200 ml Krebs Henseleit hydroxy-ethyl-starch buffer solution (KHB) by a roller pump at a constant volume inflow of 150 ml/min in a closed recirculatory system with an open reservoir to collect outflow from the left atrium. Pulmonary embolism by circulating particles or air was prevented by a 40-μm filter and a bubbletrap placed right before the pulmonary artery catheter. The lungs were ventilated with 4% CO²in air (frequency, 25 breaths/min; tidal volume, 30 ml; positive end-expiratory pressure, 0.5–1.0 cm H²O) and, to avoid atelectasis formation, intermittently expanded by increasing the expiratory pressure up to 3 cm H²O for three inspirations. The mean pulmonary arterial pressure (mpap) and airway pressures were continuously recorded via  Statham strain-gauge transducers. Because oxygenation of the lung by ventilation was satisfactory and previous experiments did not show any hypoxic cellular damage, 20no additional oxygenator unit in the perfusion circuit was required. Because of a constant perfusion flow, alterations of perfusion pressure directly reflect alterations of pulmonary vascular resistance. Samples of perfusate were taken intermittently from a catheter that collects the effluent from the pulmonary veins, for measurements of p  H, oxygen and carbon dioxide partial pressure, oxygen saturation (blood gas analysis system 288; Ciba Corning, Fernwald, Germany), oncotic pressure (Onkometer BMT 921, Dr. Karl Thomae GmbH, Berlin, Germany), and determination of TXB²concentrations. The lungs were initially perfused with KHB solution using low flow rates in an open circulatory system to remove remaining blood from the vascular bed. The perfusion fluid was then exchanged for fresh buffer via  two separate perfusion circuits 2 min after the beginning of the extracorporeal circulation and 15 min later, after the flow was increased to 150 ml/min. After another 30-min steady-state period, these lungs had a constant mean mpap of 8–9 mmHg (zero-referenced at the hilum). The only lungs selected for the study were those that showed a homogenous white appearance with no signs of hemostasis or edema formation and that had no changes in weight during the steady-state period. In previous experiments, perfusion with KHB has been documented to maintain integrity of the microcirculation for more than 5 h in our model, which was assessed by measurements of mpap and weight gain, by biochemical analysis (lactate dehydrogenase, arachidonic acid metabolites, histamine), and by ultrastructural studies. 20 

Male Sprague Dawley rats (250–300 g) were housed with free access to food and water. The main pancreatic duct was cannulated under ether anesthesia, and edematous pancreatitis was induced by intravenous infusion of cerulein (5 μg · kg⁻¹· h⁻¹) for 6 h. 21Pancreatic juice was collected on ice for several hours as previously described 22and immediately frozen, lyophilized, and stored at −70°C.

A specific antirat-PAP antibody 23was prepared according to the method described by Iovanna et al.  24and was covalently bound to a solid matrix (Affi-Gel 10 gel; Bio-Rad, Ivry sur Seine, France) according to the manufacturers’ specifications. The column was equilibrated with 2-[N -morpholino] ethane sulfonic acid buffer (20 mM;p  H 6.5) with 150 mM NaCl. Lyophilized pancreatic juice was resuspended in 2-[N -morpholino] ethane sulfonic acid buffer (20 mM;p  H 6.5), 150 mM NaCl, 5 mM benzamidine, 2 mM phenylmethylsulfonyl fluoride, and 5 mM N α-benzoyl-L-arginin. After centrifugation (3,000 g  for 15 min), the supernatant was applied to the affinity gel column and recirculated overnight at 4°C. Thereafter, the column was washed with 2-[N -morpholino] ethane sulfonic acid buffer to elute unbound contaminants. The bound protein was eluted with glycine buffer (200 mM;p  H 2.8), dialyzed against phosphate-buffered saline buffer, and concentrated by centrifugation with Centrifugal Ultrafree-20-Kit (Millipore; Saint-Quentin, France).

Protein concentration was measured according to Bradford. 25Part of the protein solution was submitted to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (15% acrylamide; 0.5% bisacrylamide), which was performed according to standard procedures. 26The gels were either stained by a silver-staining method, 27or the samples were blotted on nitrocellulose membranes, and Western blot analysis was performed with a specific antirat-PAP antibody. 23,24,26Silver staining revealed a single protein band corresponding to the molecular weight of PAP, which crossreacts with the antirat-PAP antibody during Western blot analysis, indicating that the purified protein was indeed PAP.

Thirty lung preparations were randomly assigned to the different groups. After a 30-min equilibration period, the first perfusate sample was drawn for measurements of baseline values. In pilot experiments, the effects of different concentrations of PAP (260 μg/l, n = 3; 500 μg/l, n = 3; 1,000 μg/l, n = 3) were investigated. Therefore, only PAP was administered after the steady-state phase, and mpap and weight gain of the lungs were monitored for 150 min. Samples of perfusate were taken at 5, 10, 30, 60, 120, 150 min to investigate TXB²release and to analyze blood gases.

As in pancreatitis-induced lung injury, which is likely to be leukocyte-mediated via  inflammatory receptor stimulation, neutrophil activation was achieved in six lung preparations without further interventions. For this purpose, the chemotactic peptide fMLP was added after a 30-min steady-state period, yielding a final concentration of 10⁻⁶M in the perfusion fluid, and monitoring was performed as described for the pilot experiments (control group). To study protective effects of PAP on polymorphonuclear neutrophil leukocyte–mediated lung injury, in the experimental groups PAP was added to the perfusion fluid in two different concentrations (260 μg/l, n = 9; or 500 μg/l, n = 6) 15 min before fMLP administration (10⁻⁶M). These concentrations have been reported in moderate to severe pancreatitis. 11The applied time schedule was chosen to guarantee a uniform distribution of PAP in the whole lung before inflammatory activation, which induces transient perfusion inhomogenities. Monitoring of mpap, weight gain, TXB², and blood gases was performed similar to the other groups.

Thromboxane B²was assayed from 100 μl of recirculating KHB as stable hydrolysis product of TXA²by an enzyme immunoassay (TiterZyme TXB²EIA; PerSeptive Diagnostics Inc., Framingham, MA). The immunoassay and the photometric measurement (AR 2001 photometer; Anthos Labtech Instruments, Krefeld, Germany) were performed according to the manufacturers specifications. The crossreactivity of TXB²-antiserum with 2,3-Dinor-TXB²was 55.8%, 1.5% with 11-Dehydro TXB², 1.0% with prostaglandin B², and 0.5% with prostaglandin D². Other eicosanoids exert cross reactivity less than 0.1%.

Because the lungs in the current study were perfused in a closed circuit, the inactive metabolite TXB²accumulated over the observation period in the perfusate. Thus, TXB²release was calculated per minute as the difference of TXB²to the previous value divided by the time in minutes.

A cell- and plasma-free perfusion medium was used to avoid the complex interactions with different circulating cells, which may mask direct effects on vascular tone and mediator release. The perfusate consisted of a Krebs-Henseleit buffer solution with a colloid oncotic pressure between 23 and 25 mmHg, yielding the following final concentrations: Na⁺, 138 mM; K⁺, 4.5 mM; Mg⁺⁺, 1.33 mM; Cl⁻, 135 mM; Ca⁺⁺, 2.38 mM; glucose, 12 mM; HCO³⁻, 12 mM. The osmolality was approximately 330 mOsm/kg (Mikro-Osmometer; Roebling Meßtechnik, Berlin, Germany). The p  H of the buffer solution was adjusted to 7.4 with 1 M NaHCO³. Effects caused by endotoxin contamination of the plasma-free perfusate can be excluded in our model, as assessed in previous experiments. 28No hemodynamic reactions, thromboxane generation, or histamine release was observed after endotoxin addition to the perfusate in the absence of plasma complement components. 28 

N -formyl-Met-Leu-Phe was purchased from Sigma Chemicals (Deisenhofen, Germany), dissolved in dimethyl sulfoxide, and stored deep-frozen until use. Final concentrations of fMLP in the pulmonary circuit were 10⁻⁶M.

Data are presented as mean ± SD. Differences between groups were tested by one-way analysis of variance followed by a Student Newman-Keuls multiple-comparison procedure. Significance was accepted at P < 0.05.

This study was approved by the Animal Subject Protection Committee of the local government. The care and handling of animals were in accordance with the principles expressed in the Helsinki Declaration.

After an equilibration period of 30 min, baseline values of mpap (8 ± 2 mmHg) and airway pressure (6–8 cm H²O) were similar in all groups and in agreement with previous studies reported by our group. 18Because effects of PAP on the pulmonary vasculature are unknown, administration of PAP was investigated to determine its properties in our lung preparation. During perfusion with any concentration of PAP (260 μg/l, n = 3; 500 μg/l, n = 3; 1,000 μg/l, n = 3), no changes in mpap, weight gain, or TXB²release were observed, and the lung preparations were stable until the end of the 150-min observation period. The leukocyte activation with fMLP in the lungs resulted in an acute increase of mpap up to 26 ± 13 mmHg (P < 0.01) within 5 min (fig. 1). Baseline levels of mpap were regained 30 min after fMLP addition and were then maintained for 120 min. A dose-dependent effect on mpap was observed after treatment with PAP. The pressure response was reduced significantly to 15 ± 7 mmHg (PAP 260 μg/l;P < 0.05) and 9 ± 4 mmHg (PAP 500 μg/l), respectively. Mean pulmonary artery pressure returned to baseline 10 min after fMLP and remained almost constant until the end of the experiment. fMLP induced lung weight gain of 9 ± 7 g during the 150-min observation period, whereas pretreatment with PAP in either concentration significantly (P < 0.05) prevented pulmonary edema formation (PAP 260 μg/l, 2 ± 2 g; 500 μg/l, 1 ± 2 g;fig. 2).

Analysis of perfusate samples revealed a baseline concentration of circulating TXB²of 2.63 ± 2.60 nmol/ml and an immediate increase (P < 0.005) after fMLP injection. PAP significantly suppressed TXB²formation caused by fMLP injection compared with controls. Although mean values in controls reached 1,138 ± 800 pmol · ml⁻¹· min⁻¹, TXB²detected in the PAP groups were 200 ± 220 pmol · ml⁻¹· min⁻¹(P < 0.01; PAP 260 μg/l) and 285 ± 70 pmol · ml⁻¹· min⁻¹(P < 0.05; PAP 500 μg/l) 5 min after fMLP injection (fig. 3).

Oxygen saturation, oxygen and carbon dioxide partial pressures, and p  H and base excess did not significantly change throughout the observation period in all experiments.

Pulmonary dysfunction such as pulmonary edema and hypoxemia is one of the main systemic complications in acute pancreatitis. 1,6Experimental studies provided evidence that pancreatitis induces cytokines that activate granulocytes in the lung vasculature, initiating pulmonary damage 15by elastase and phospholipase A²release. 29Despite intensive research efforts, until now the effects of potential modifiers of this inflammatory response, such as heat shock proteins and PAP, are not well understood. Recently, characteristic alterations of protein synthesis at translational and transcriptional levels were detected during acute pancreatitis. 8The most prominent finding was the overexpression of an unique c-lectin protein in the pancreas of animals 9and humans. 10This protein is absent in the normal gland but synthesized in high amounts in the inflamed pancreas. 12During inflammation, evidence accumulated that the acute-phase protein response, including PAP, exerts beneficial effects in different species and in distinct experimental settings. 14After induction of acute pancreatitis in rats, PAP was first detected after 6 h. Maximum levels corresponding to a 400–1,000-fold increase of PAP were measured within 48 h. 11,30Considering that high amounts of this unique protein are expressed during acute pancreatitis correlating with the severity of disease, 11but also correlating with increased survival in rats, 14our interest was focused on whether PAP might show organ protective properties in a model of pancreatitis-associated lung failure. Because alterations in vascular tone and a capillary leak are hallmarks of acute respiratory distress syndrome, the study was designed to investigate the pathophysiologic effects of this protein on fMLP-induced leukocyte activation and on subsequent pulmonary vascular resistance and edema formation. For this purpose, highly purified PAP from rat pancreatic juice was added to the perfusion fluid of the isolated lung. The applied dosage in our model was comparable to serum levels measured in patients with mild to severe pancreatitis. 11To guarantee homologous distribution of PAP in the lung, PAP was administered before the inflammatory fMLP stimulus.

Although the isolated-cell free perfused rabbit lung is a relatively simple model that allows the investigation of PAP effects under standardized conditions, there are particular limitations. It does not allow analysis of interactions with other organ systems or between the endothelium and blood components. Moreover, the observation period is limited to approximately 6 h. However, this experimental set-up allows selective measurements of alterations in pulmonary vascular resistance and permeability as well as the release of mediators into the perfusate. Therefore, conclusions on pathophysiologic mechanisms can be drawn from this model. 18–20It has been shown that granulocyte activation by proinflammatory cytokines plays a crucial role in pulmonary damage during pancreatitis. 4,5Thus, activation of granulocytes by fMLP in the lung vasculature was chosen in the present study. Although fMLP-induced lung injury may reproduce some features of injury produced by pancreatitis, inflammation in vivo  is considerably more complex, characterized by a mounting cascade of a variety of mediators released from a variety of inflammatory cells. These additional factors may modulate the actual effect of PAP during clinical pancreatitis. The time course of pulmonary artery pressure, with maximum vasoconstriction after 5 min and decreasing values until baseline at 30 min, is in agreement with that of other investigators 31using fMLP in isolated lungs.

The current data on PAP-induced effects on lung vascular tone and edema formation demonstrate protective properties of PAP on leukocyte-induced pulmonary injury. Pulmonary artery pressure (fig. 1) as well as edema formation (fig. 2) in the lung were significantly suppressed after pretreatment with PAP, whereas application of PAP without leukocyte activation did not affect any measured parameter. In seeking to understand the protective effects of PAP concerning the acute pressure increase after polymorphonuclear neutrophil leukocyte activation, the role of TXA²as a potent vasoconstrictive mediator in the pulmonary circulation 32,33was investigated in the presence and absence of PAP. The generation and release of this mediator is induced by a variety of factors, including tissue damage, allergic reaction, infection, and other inflammatory diseases. TXA²is supposed to be the predominant mediator of arachidonic acid–induced pressure responses in rabbit lungs, surpassing the vasodilatory properties of simultaneously generated prostacyclin. This was confirmed in our model in previous studies. 34In the current study, the release of TXA²into the perfusate corresponded with an increase in pulmonary artery pressure in all experimental set-ups. Hence, PAP dose-dependently suppressed TXA²generation, which was associated with the modulation of the pulmonary pressure response (fig. 3).

In the present study, PAP was obtained from rats by affinity chromatography, which is highly selective for PAP, and its purity was checked carefully, 23,25–27indicating that the isolated protein indeed was PAP. Thus, effects caused by impurities can be excluded, and the present data precisely reflects properties of PAP. Homology of human and rat PAP has been described 35in terms of size, 71% amino acid identity, and the six half-cystines being in identical positions. Because rabbit PAP has not been sequenced, assumptions about the structure of rabbit PAP remain speculative. Nevertheless, the data obtained in this xeno-model indicate protective effects of the investigated protein without any signs of foreign protein reaction, which would evoke alterations in vascular tone and permeability in isolated lungs. 36Because PAP has been described as a nonenzymatic protein without protease activity, its effects on fMLP-induced pulmonary damage are very unlikely to be caused by hydrolysis of fMLP, which would cause insufficient leukocyte activation. The distinct molecular mechanisms of action cannot be answered from the current study design.

Antiapoptotic properties of PAP have been reported 37that protect AR4–2J cells from oxygen radical–induced cell injury. Although reactive oxygen species play a crucial role in pancreatitis-induced pulmonary damage and were released in our model after fMLP stimulation, protection from apoptosis might contribute to the observed effects. Whether the observed protective effects of PAP on acute lung injury play a significant role in human acute respiratory distress syndrome after acute pancreatitis remains to be elucidated.

In summary, PAP prevents polymorphonuclear neutrophil leukocyte–mediated pulmonary vasoconstriction and edema formation in the isolated rabbit lung in concentrations found during pancreatitis. These findings point toward a protective effect of PAP on pancreatitis-associated lung injury.

The authors thank K. Salomon, J. Schulte, and A. Tapper (Institute of Anesthesiology, University Hospital Mannheim, Mannheim, Germany); J. Christophel (Center of Medical Research, University Hospital of Mannheim, Mannheim, Germany); and M. A. Devaux (INSERM U315, Marseille, France) for excellent technical assistance.