Nitrosative Stress and Myocardial Sarcoplasmic Endoreticular Calcium Adenosine Triphosphatase Subtype 2a Activity after Lung Resection in Swine

Under protocols approved by the Institutional Animal Care and Use Committee at the Weill Cornell Medical College, New York, New York, 16 swine (30–42 kg in weight) were used for the study. Seven animals underwent removal of the left upper lung lobe during single lung ventilation to facilitate lung collapse, 4 underwent thoracotomy without lung collapse or lobectomy (sham procedure), and the remaining 5 were nonoperated controls. Both the lobectomy and sham groups were studied on the third postoperative day. This time frame was chosen to assure that the acute effects of surgery had dissipated and that all animals had resumed normal activity.

After induction of anesthesia with intramuscular tiletamine–zolazepam (4.4 mg/kg) and xylazine (2.2 mg/kg), the trachea was intubated, the lungs were ventilated with oxygen and isoflurane, and, for the lobectomy animals, a balloon-tipped catheter was inserted into the left mainstem bronchus under bronchoscopic guidance. When filled with sufficient air, the balloon catheter prevented ventilation of the operative lung during the resection. With animals in right lateral decubitus position and the bronchial-blocking balloon filled to prevent ventilation of the left lung in the lobectomy group, a thoracotomy was performed. For lobectomy animals, anatomic resection of the left upper lobe was performed using vascular and bronchial staplers. For sham animals, no additional surgery was performed in the chest, but the cavity was left open for 40 min, the approximate time period required for lobectomy. For both groups, a 20-French chest tube was then placed and connected to a Heimlich valve. Perioperative analgesia was provided with intrathecal and intravenous morphine in combination with intercostal nerve blockade. Postoperatively, analgesia was maintained with transcutaneous fentanyl and subcutaneous ketoprofen.

Animals were anesthetized as described above (Survival Lobectomy and Sham section), the trachea was intubated, and the lungs were ventilated with oxygen and isoflurane to maintain end-tidal isoflurane concentration at 1.3% and carbon dioxide at 32–34 mmHg. In addition to the animals that had undergone lobectomy or sham thoracotomy, the nonoperated control animals were prepared in an identical fashion. A femoral arterial catheter was inserted percutaneously for blood pressure monitoring, and a pulmonary artery catheter was inserted via  cut-down of the right external jugular vein. A median sternotomy was performed, and the heart was suspended in a pericardial cradle. At the apex, a 5-French micromanometer catheter was inserted via  a puncture into the left ventricle for pressure measurement and secured with a purse string suture. In the left ventricular free wall, a pair of 2-mm ultrasonic crystals were inserted into the myocardium approximately 15 mm apart for measurement of left ventricular segment length as a surrogate for volume. Hemodynamic variables were then recorded with left ventricular contractility expressed as peak +dP/dt normalized to left ventricular end-diastolic segment length over the variation in preload induced by acute occlusion of the inferior vena cava. For all animals, stroke volume and cardiac output data were normalized to body surface area20and were expressed as stroke volume index and cardiac index, respectively. After in situ  data collection, the heart was arrested with iced potassium chloride solution and rapidly removed. Myocardial samples were snap frozen in liquid nitrogen and either used for preparation of SR vesicles (see Force–Frequency Relations in Isometric Trabeculae section below) or stored at −80°C until analysis. In addition, for control and lobectomy animals, free wall tissue was also placed into ice-cold, oxygenated Krebs-Ringer’s solution for isolation of myocardial trabeculae (small muscle strips 1–1.5 mm in diameter, approximately 6 mm in length) and subsequent assessment of contractile reserve (see Force–Frequency Relations in Isometric Trabeculae section below).

For the lobectomy and sham groups, arterial blood samples were drawn into EDTA tubes before opening, and just after closing, the chest during the survival procedure, and again before beginning the organ harvest. For the control animals, blood samples were obtained only at organ harvest. Using the Malondialdehyde Assay Kit (Northwest Life Science Specialties, Vancouver, WA), plasma was added to individual wells and processed according to the manufacturer’s specifications. The resulting malondialdehyde levels reflect peroxidation reactions initiated by oxygen radicals and therefore represent a stable index of oxidative stress.21In that inflammation can also elicit local oxidative stress via  activation of leukocytes and myeloperoxidase is a reflection of this activation,22myocardial samples obtained at sacrifice were analyzed for myeloperoxidase content using a commercial kit (Northwest Life Science Specialties) after tissue homogenization in the recommended lysis buffer.

Although the profoundly reactive peroxynitrite molecule has an extraordinarily short half-life, the resulting modification of tyrosine residues within proteins (addition of a nitro group to position 3 of the phenyl ring) represents a stable end-product that can be used to index peroxynitrite production.4,23After homogenization in 10 ml/g wet weight of tissue in a buffer consisting of 50 mm Tris, 150 mm NaCl, 0.1 mm EDTA, and 20 mm CHAPS (pH 7.4), the samples were centrifuged at 100,000g  for 1 h and the supernatants were removed to fresh tubes. A 100-μl aliquot of the tissue homogenate was then subjected to proteolytic digestion before the extent of protein incorporated 3-nitrotyrosine was quantified. Proteinase K (48U/ml buffer) was added to the tissue homogenate and incubated for 8 h at 55°C. Samples were then allowed to cool to room temperature before addition of ice-cold acetonitrile (3:1 vol/vol), followed by vortexing, cooling on ice for 5 min, and centrifugation at 12,000g  at 4°C for 15 min. The resulting supernatants were removed to fresh tubes, dried in a speed vacuum, and resuspended in 400 μl mobile phase containing 90 mm sodium acetate, 35 mm citric acid, 130 μm EDTA, and 460 μm sodium octane sulfonate (pH 4.35). Samples were then processed using an isocratic high-performance liquid chromatography system with multichannel electrochemical CoulArray (ESA Inc., Chelmsford, MA) detection to effectively resolve 3-nitrotyrosine from background species with a 100-mm C-18 column (Microsorb-MV; Varian Inc., Palo Alto, CA) running the mobile phase at a flow rate of 0.75 ml/min and the temperature at 30°C. The optimum potential for detection of 3-nitrotyrosine is +800 mV. For maximum selectivity of the system for 3-nitrotyrosine, two additional electrodes were set either side of the optimum potential for 3-nitrotyrosine, +700 mV and +900 mV, respectively. Further confirmation of the elution of 3-nitrotyrosine was established by addition of 10 mm sodium hydrosulfite to nitrotyrosine. This treatment chemically reduces 3-nitro- to 3-amino-tyrosine, silencing the electrochemical signal.

Enriched cardiac SR samples were prepared according to Tupling et al.  24with modification. Briefly, ventricular tissue in liquid nitrogen was ground with mortar and pestle into a fine powder and diluted 1:5 (wt/vol) in ice-cold buffer containing 5 mm HEPES, 250 mm sucrose, 0.2% sodium azide, and 0.2 mm phenylmethylsulfonyl fluoride (pH 7.5). Samples were homogenized and then centrifuged in ordered steps. Final pellets were resuspended in homogenization buffer plus 50 mm l-histidine, then frozen in liquid nitrogen and stored at −80°C for later analysis. Oxalate-supported Ca²⁺uptake in enriched SR samples was subsequently measured using the Ca²⁺fluorescent dye Indo-1 Pentapotassium Salt (Calbiochem, La Jolla, CA). With methods adapted from Tupling et al.  24and O’Brien et al. ,25,26measurements were obtained with a luminescence spectrometer (dual emission monochromators) and the FL WinLab software (Perkin Elmer, Beaconsfield, England). The samples were excited at 360 nm with a bandpass of 2.5 nm. The emission monochromators were set at 400 nm (for Ca²⁺bound) and 485 nm (for Ca²⁺free) with 3.0-nm band passes. Before beginning, the protein concentration of SR isolates was quantified using a Bradford assay. With temperature maintained at 37°C, indo-1 [1.5 μm]^fand CaCl²[3.0 μm]^fwere first added to the cuvette containing uptake buffer (200 mm KCl, 20 mm HEPES, 10 mm NaN³, 5 μm N ,N ,N ′,N ′-tetrakis(2-pyridyl-methyl) ethylenediamine, 5 mm oxalate, and 15 mm MgCl², pH 7.0). The SR sample was then added. After 2 min, adenosine triphosphate (ATP) [1.5 mm]^fwas added to initiate the reaction. When uptake reached plateau, 50 mm EGTA and then 100 mm CaCl²were added to calibrate indo-1 intensities for use in calculation of Ca²⁺concentration.25SERCA2a dependence of the preparation was confirmed by preincubation with 1.25 μm thapsigargin, a SERCA inhibitor, which abolished calcium uptake. In addition, samples were analyzed in the presence of 300 mm ryanodine to determine whether SR calcium release via  the ryanodine receptor represented a confounding variable.

To determine whether exposure to peroxynitrite specifically alters calcium uptake, SR vesicles from control hearts were exposed to varying concentrations of peroxynitrite for 3 min in the dark before measurement of calcium uptake. In that peroxynitrite concentration within a stock solution can vary widely over time, the actual concentration present was assayed before each experiment using the method of Beckman et al.  27Briefly, an aliquot of the peroxynitrite solution was diluted with 1.2 m NaOH, absorbance was measured at 302 nm, and the concentration was calculated using the extinction coefficient, ϵ= 1,670 m ⁻¹· cm⁻¹. As a negative control, an aliquot of peroxynitrite was first prepared in 0.5 m phosphate buffer (pH 7.4). Under these neutral pH conditions, peroxynitrite degrades, rapidly yielding an absorbance that is the same as that of buffer alone. Finally, to determine whether SERCA2a activity is sensitive to varying concentrations of ATP and whether the response is altered by preincubation with peroxynitrite and/or consistent among groups, SR calcium uptake measurements were performed on the same samples using different ATP concentrations to initiate the reaction.

Myocardial samples were homogenized in ice-cold buffer containing 20 mm Na-HEPES (pH 7.4), 4.0 mm EGTA, 1.0 mm dithiothreitol, and EDTA-free protease inhibitors (Boehringer-Mannheim, Indianapolis, IN) and were then centrifuged, and protein content of the supernatant was determined by Bradford assay. Samples were then subjected to 10% SDS-PAGE, transferred to nitrocellulose, and blocked with 2.5% nonfat dry milk in phosphate buffered saline with Tween 20 (PBS-T) buffer. After washing with PBS-T, blots were incubated with primary antibody for 2 h at room temperature. Primary antibody was diluted in 2.5% nonfat dry milk in PBS-T buffer and included mouse monoclonal antibodies against SERCA2a (Calbiochem), and actin (Oncogene Research Products, La Jolla, CA). Protein bands were analyzed by laser densitometry using actin as an internal standard.

After homogenization as noted above and determination of protein content, samples were subjected to 10–20% SDS-PAGE, transferred to nitrocellulose, and blocked with 2.5% nonfat dry milk in PBS-T buffer. After washing with PBS-T, blots were incubated with primary antibodies for total phospholamban as well as phospho-Ser16 phospholamban (P-Ser16 PLB), phospho-Thr17 phospholamban (P-Thr17 PLB), and actin for 2 h at room temperature. After application of secondary antibody (all anti-rabbit immunoglobulin G) and development with enhanced chemiluminescence, protein bands, reflecting both the pentameric and monomeric forms, were analyzed by laser densitometry using actin as an internal standard. P-Ser16 PLB and P-Thr17 PLB were expressed as a fraction of total phospholamban.

Excised trabeculae were allowed to equilibrate for at least 1 h at low tension (< 50 mg) at 37°C. Aerated Krebs-Henseleit solution gassed with 95% O²–5% CO²was replaced every 30 min. Electrical field stimulation was continuously applied using a Grass S48 stimulator and a Grass SIU5 stimulus isolation unit (both Astro-Med Inc., Grass Technology Product Group, West Warwick, RI) (1.5 Hz, 90 V, 5 ms). After completion of the initial equilibration period, tissues were slowly stretched to 600 mg and then further progressively stretched to the length of submaximal force generation (approximately 90% Emax). After an additional reequilibration of 30 min, the stimulation frequency was increased every 2 min at 0.5-Hz increments to a maximum of 3.0 Hz (180 bpm). For each preparation, isometric myocardial force generation was assessed using Powerlab/8SP (AD Instruments, Colorado Springs, CO) and referenced to the baseline level at 1.5 Hz.

All data are presented as mean ± SE. For hemodynamic and biochemical variables, comparisons between groups were made with one-way analysis of variance. For comparison of different concentrations of ATP on SR calcium uptake in myocardial samples from the same animal, a paired t  test was applied. For isolated myocardial trabeculae data, the contractile force of control and lobectomy tissues was compared by nonpaired t  test. For all analyses, P ≤ 0.05 was considered significant.

Hemodynamic variables obtained with the chest open and identical end-tidal isoflurane and carbon dioxide levels are shown in table 1. There were no differences in hemodynamic variables between groups, although a trend toward higher mean pulmonary artery pressure and pulmonary vascular resistance was evident in lobectomy animals. Under these “resting” conditions, there were no differences in inotropy among the groups.

As shown in figure 1, during the survival lobectomy and sham procedures, malondialdehyde levels were the same before opening the chest, but twice as high in lobectomy animals after chest closure. At the time of organ harvest, however, malondialdehyde levels in the lobectomy group were not different than in either sham or control animals. Similarly, at the time of organ harvest, there were no differences in myocardial myeloperoxidase content among the groups, indicating that myocardial inflammation and local leukocyte activation was not a source of ongoing oxidative stress within the heart.

Despite the lack of biochemical evidence for ongoing systemic and/or myocardial oxidative stress, as shown in figure 2, the myocardial 3-nitrotyrosine content was more than twofold higher in lobectomy animals relative to both control and sham animals, indicating substantial generation of peroxynitrite within the tissue. This increased level of myocardial 3-nitrotyrosine was associated with a nearly 50% lower rate of SERCA2a-mediated calcium uptake in lobectomy animals relative to both control and sham animals (fig. 3). For all preparations, calcium uptake was enhanced to the same degree (approximately 7%) by preincubation with 300 mm ryanodine (data not shown), indicating that decreased uptake in the lobectomy animals was a result of altered SERCA2a activity and not due to increased release of calcium via  the ryanodine receptor. To determine whether exogenous peroxynitrite similarly alters SERCA2a function in this preparation, control samples were incubated with peroxynitrite beforemeasurement of calcium uptake.1As shown in figure 4, preincubation with 100 μm peroxynitrite reduced SERCA2a-mediated SR calcium uptake by half. However, this effect was partially reversed by increasing the ATP concentration, a finding suggestive of a change in SERCA2a affinity for ATP. Finally, to determine whether SERCA2a activity in lobectomy preparations could also be restored by increased ATP concentrations, SR calcium uptake was repeated in the presence of increasingly greater concentrations of ATP. As shown in figure 4, doubling the ATP concentration increased SR calcium uptake in lobectomy preparations but had no effect in either control or sham myocardium.

Decreased SR calcium uptake in lobectomy animals was not associated with a decrease in SERCA2a protein (fig. 5). To evaluate whether changes in phospholamban expression or phosphorylation status (dephosphorylated phospholamban inhibits SERCA2a activity) contributed to the postlobectomy decrease in SR calcium uptake, blotting with both nonspecific antiphospholamban and phosphospecific (Ser16 and Thr17) antibodies was performed. As shown in figure 6, there were no discernible differences in total phospholamban or P-Ser16 PLB and P-Thr17 PLB among the groups.

As shown in figure 7, trabeculae from control animals exhibited a rate-related increase in contractile force (positive force–frequency relation), indicating a “recruitable reserve” in SERCA2a function. In contrast, those from lobectomy animals exhibited decreased force with increasing stimulation frequency (negative force–frequency relation), suggesting that with increased rate, SERCA2a could not clear calcium from the cytosol fast enough to optimally load the SR for the subsequent systole.

As a consequence of the myocardial oxidative stress induced by aging and a variety of chronic diseases, superoxide and nitric oxide interact within the heart to produce peroxynitrite that, in turn, nitrates tyrosine residues within many proteins potentially altering function.3,23Whether acute perioperative oxidative stress such as that elicited by pulmonary resection during intentional lung collapse can elicit similar nitrosative effects on the myocardium is unknown. However, given the profound reactivity of peroxynitrite and the fact that tyrosine nitration is generally considered a permanent protein modification,3the possibility exists that if a discreet oxidative challenge is of sufficient magnitude, a persistent effect on vulnerable aspects of myocardial function may occur.

Although global significance of nitrosative protein modifications in the heart remains uncertain,28exposure to peroxynitrite clearly reduces activity of SERCA2a, a key determinant of calcium cycling within the myocyte.1,2,4,9,10,23The current study was therefore designed to use SERCA2a activity as a functional endpoint in testing the general hypothesis that pulmonary resection during selective lung collapse elicits myocardial peroxynitrite generation and postoperative impairment of cardiac protein function. Results of the study support this hypothesis in that they confirm the presence of substantial myocardial peroxynitrite generation that is associated with impaired SERCA2a activity unrelated to protein expression or phosphorylation state of the SERCA2a regulatory protein phospholamban.

The concept of lung resection eliciting a short-term systemic oxidative stress is well described. Previous investigators have reported exhaled peroxide along with large increases in plasma markers for oxidative stress that occur during lobectomy, presumably due to intentional lung collapse followed by hyperoxic reexpansion.15–18Recent data have characterized this response (quantified as plasma malondialdehyde) as peaking shortly after reexpansion of the remaining operative lung and then declining to a level below baseline over the next 8 h.16,17Although previous studies have attempted to link the magnitude of this short-term increase in plasma malondialdehyde to complications after lobectomy,16there are no clinical data indicating a specific physiologic effect of the acute oxidative stress associated with lung resection. We previously reported experimental data demonstrating that after lobectomy during intentional lung collapse, tetrahydrobiopterin, a critical cofactor for nitric oxide synthase, becomes oxidized to dihydrobiopterin within both the remaining operative lung and the nonoperated contralateral lung.29Functionally, this oxidation of tetrahydrobiopterin uncouples components of the nitric oxide synthase complex, causing the enzyme to produce superoxide instead of nitric oxide. The end result is not only endothelial dysfunction (inability to produce nitric oxide), but also the possibility of further local oxidative stress via  a “feed forward” mechanism, i.e. , ongoing superoxide formation leads to more oxidation of tetrahydrobiopterin.

For the current study, we chose to focus downstream on myocardial 3-nitrotyrosine as a marker for peroxynitrite generation within the heart in response to oxidative stress. Consistent with clinical studies, our data indicate an acute increase in malondialdehyde after lung resection, but not sham thoracotomy, that is no longer present by the third postoperative day. However, despite normalization of plasma malondialdehyde levels, 72 h after lung resection, myocardial 3-nitrotyrosine incorporation was more than twofold higher than that measured in both nonoperated controls and animals that had undergone sham thoracotomy. These data strongly suggest that the lung resection procedure elicited sufficient myocardial oxidative stress to trigger substantial peroxynitrite generation within the heart. It is unclear, however, whether this response reflects events that occurred intraoperatively or ongoing postoperative processes. Nonetheless, myocardial levels of myeloperoxidase were found to be the same in control, lobectomy, and sham animals, suggesting that ongoing myocardial inflammation after lobectomy was not a major component of the response.

Previous studies have indicated that cardiac SERCA2a activity is reduced by aging and pathologic conditions such as dilated myopathy and diabetes.1,10,11,30While recent data have raised the prospect that SERCA2a nitration may be a contributing factor,1,23diminished SERCA2a activity has largely been attributed to reduced protein expression and/or a shift in the expression or phosphorylation state of the SERCA2a regulatory protein phospholamban. For example, it has been speculated that because unphosphorylated phospholamban elicits tonic inhibition of SERCA2a, any change in SERCA2a expression must be interpreted in light of simultaneous changes in phospholamban, i.e. , the SERCA2a/phospholamban ratio has functional significance.31Alternatively, because phosphorylation of phospholamban removes inhibitory effects of the protein on SERCA2a function, alterations in protein phosphatase activity may need to be considered because this effect can modulate SERCA2a activity by affecting the rate at which phospholamban is dephosphorylated.30In the current study, neither a shift in the SERCA2a/phospholamban ratio or a change in phosphatase activity seemed to be factors because we did not observe a difference among groups in regard to expression of SERCA2a or phospholamban, or phospholamban phosphorylation at either Ser16 and Thr17. Accordingly, although our data do not specifically demonstrate a direct SERCA2a modification by peroxynitrite, this prospect is supported by the marked decline in SERCA2a activity independent of any changes in the expression of SERCA2a or phospholamban, or phospholamban phosphorylation. Furthermore, the finding that SERCA2a activity in lobectomy animals and control samples treated with peroxynitrite, but not untreated controls, can be enhanced by increased ATP suggests a change in ATP affinity, a finding seemingly consistent with reported steric effects of SERCA2a modification by peroxynitrite.4 

In that rate-related changes in myocardial inotropy are regarded in some ways as a bioassay for SERCA2a,19,32the observation that myocardial trabeculae isolated from control animals exhibited augmented contractile force when stimulated with increasing frequency (positive force–frequency relation) whereas those harvested from lobectomy animals exhibited decreased contractile force (negative force–frequency relation) is consistent with impaired postoperative SERCA2a activity and calcium cycling. Nonetheless, there were no in situ  differences in ventricular inotropy. Although seemingly at odds with the isolated tissue responses, these observations are, in fact, in accord with previous data demonstrating that decreased SERCA2a activity has a greater impact upon inotropic reserve than resting function in myocardial trabeculae.32It is possible, however, that postoperative changes in mechanical reserve could have been modulated, at least in part, by other reported effects of oxidative/nitrosative stress3or as a result of some persistent, ill-defined aspect of anesthesia and/or surgery. Unfortunately, trabeculae were not harvested from the sham animals, so it remains unknown whether the force–frequency relation of trabeculae isolated from this group would be positive or negative. Whatever the underlying mechanism behind diminished rate-related contractile reserve in the lobectomy animals, our observations are consistent with a diminished myocardial capacity to compensate for postoperative stress19,32and suggest that although resting cardiac function was not influenced by lung resection, the response to stimuli such as fever, exercise, or supraventricular tachycardia may well have been altered.

Results of the study must be interpreted in the context of certain limitations. First, although both myocardial peroxynitrite generation and SERCA2a inhibition are evident on the third postoperative day, the data do not provide evidence for concomitant onset or duration, or establish a temporal relation between stimulus and maximal response. Second, the study does not define the precise functional role of myocardial peroxynitrite formation. Although there are considerable, compelling data in support of SERCA2a modification and inhibition by peroxynitrite, whether interruption of peroxynitrite generation by pretreatment with an antioxidant and/or blockade of nitric oxide synthesis will alter the observed effects on both SERCA2a activity and the myocardial force–frequency relation remains to be determined. Finally, the study was not designed to determine the independent effects of anesthesia, relative hyperoxia, surgical trauma, and single-lung ventilation on myocardial peroxynitrite generation and/or SERCA2a inhibition. Nonetheless, the finding that 3-nitrotyrosine and SERCA2a activity data from the sham group were virtually identical to those from nonoperated controls, and distinctly different from the lobectomy animals, suggests that anesthesia, hyperoxia, and thoracotomy alone did not have any effect. Although these data must be considered in light of the relatively small sample size, analysis of variance of 3-nitrotyrosine and SERCA2a activity data demonstrated power values of 0.8 and 1.0, respectively. Accordingly, the conclusion that higher myocardial 3-nitrotyrosine incorporation and lower SERCA2a activity were primarily the result of the lung resection procedure seems valid.

In summary, results of the study indicate that lung resection is associated with significant left ventricular nitrosative stress and impaired myocardial calcium cycling. The data support the probability that acute oxidative/nitrosative stress can alter SERCA2a activity independent of any influence on protein expression or phospholamban phosphorylation and suggest a biochemical basis for perioperative changes in myocardial mechanical reserve.