Anesthesiology 2005; 102:1182–9 © 2005 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Opposing Effects of Isoflurane and Sevoflurane on Neurogenic Pulmonary Edema Development in an Animal Model Nobuhisa Kandatsu, M.D.,* Yong-Shan Nan, M.D.,† Guo-Gang Feng, M.D., Ph.D.,‡ Kimitoshi Nishiwaki, M.D., Ph.D.,§ Mitsuru Hirokawa, M.D.,* Kiyonori Ishikawa, D.D.,ʈ Toru Komatsu, M.D., Ph.D.,# Takashi Yokochi, M.D., Ph.D.,** Yasuhiro Shimada, M.D., Ph.D., F.C.C.P.,†† Naohisa Ishikawa, M.D., Ph.D.‡‡

Background: The current study was undertaken to investigate beds has also been shown to play a role in enhancing the effects of pretreatment with isoflurane and sevoflurane on increases in vascular pressure and permeability2,5 when the development of neurogenic pulmonary edema in an animal Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 model. sympathetic nerve activity is increased. Because anes- Methods: Rats were exposed to room air (control), 1.5% thetics have been thought to inhibit the activities of isoflurane, or 2.5% sevoflurane for 4 h. They were then anes- autonomic nerves, they have not been considered to thetized with intraperitoneal injections of pentobarbital so- affect NPE. However, it has been reported that some dium, and fibrinogen and thrombin were injected into the cis- inhaled anesthetic agents affect the vascular permeabil- terna magna to induce neurogenic pulmonary edema. 6 Results: Consecutive injections of fibrinogen and thrombin ity of biologic membranes to water and electrolytes ; caused increases in blood pressure, with the peak values ob- isoflurane increases the permeability of alveolar epithe- tained in the isoflurane and sevoflurane groups being lower lial cells7 and the alveolar–capillary membrane,8 and than the control values. The incidence of significant neurogenic halothane increases fluid conductance across the pulmo- pulmonary edema was 58%, 100%, and 8% in the control, isoflu- nary capillary bed,9 perhaps by interacting with oxi- rane, and sevoflurane groups, respectively. The lung water ra- ,tio, an index of severity of edema, was 4.86 ؎ 0.78, 6.15 ؎ 0.64, dants. In addition to direct actions on endothelial cells and 4.40 ؎ 0.32 in the control, isoflurane, and sevoflurane anesthetic effects on interstitial lung fluid clearance in groups, respectively. Furthermore, immunohistochemical the lungs could contribute to edema.10 Although auto- staining for vascular endothelial growth factor demonstrated nomic nerve activity may be enhanced by surgical ma- an increase of expression in the rat lungs exposed to isoflurane. 11 Treatment with an anti–vascular endothelial growth factor an- neuvers and contribute to the development of NPE, it tibody during exposure to isoflurane completely inhibited the is difficult to explain the development of NPE during effect of isoflurane to promote neurogenic pulmonary edema in anesthesia solely in relation to the actions of this model. neurotransmitters. Conclusion: Exposure to 1.5% isoflurane enhances the devel- Vasoactive substances released from vascular endothe- opment of neurogenic pulmonary edema development in this animal model, most likely via release of vascular endothelial lial cells, such as nitric oxide, bradykinin, and vascular growth factor from bronchial epithelial cells, an effect not ob- endothelial growth factor (VEGF), and agents released served with sevoflurane. by mast cells, such as histamine, can affect vascular smooth muscle and endothelial cells. In particular, VEGF THE pathogenesis of neurogenic pulmonary edema promotes endothelial cell viability, mitogenesis, chemo- (NPE) is not completely understood, but seems to be taxis, and vascular permeability. Recent in vitro vascular associated with enhanced sympathetic nerve activity.1 permeability studies have shown that it has the ability to Consistent with this concept, NPE is associated with an increase the microvascular permeability to a level 50,000 increased intravascular pressure and enhanced vascular times higher than with histamine.4,12,13 The current permeability in the pulmonary circulation, possibly me- study was undertaken to evaluate the effects of inhaled diated by sympathetic nerve neurotransmitters released anesthetics on VEGF expression and NPE development from nerve terminals, e.g., neuropeptide Y.2–4 Synergis- and to determine whether VEGF may mediate NPE de- tic interaction between neurotransmitters in vascular velopment in rats anesthetized with volatile anesthetics. Because sevoflurane has not been reported to increase vascular permeability, the effects of isoflurane on NPE * Instructor of Anesthesiology, # Professor of Anesthesiology and Chairman of the Department, ** Professor of Microbiology and Chairman of the Department, development were compared to those of sevoflurane. ‡ Research Fellow of Pharmacology, ʈ Instructor of Pharmacology, ‡‡ Professor of Pharmacology and Chairman of the Department, Aichi Medical University. † Graduate Student of Anesthesiology, § Associate Professor of Anesthesiology, †† Professor of Anesthesiology and Chairman of the Department, Nagoya Uni- versity School of Medicine, Nagoya, Japan. Materials and Methods Received from the Department of Anesthesiology, Aichi Medical University, Nagakute-cho, Aichi-gun, Aichi Prefecture, Japan. Submitted for publication Au- Animals gust 5, 2004. Accepted for publication March 3, 2005. Supported in part by grant All procedures were performed in accordance with No. 16390448 from the Japanese Ministry of Education, Science, Sports and Culture, Tokyo, Japan. the “Guiding Principles in the Care and Use of Animals in Address reprint requests to Dr. Ishikawa: Department of Pharmacology, Aichi the Field of Physiologic Sciences” published by the Phys- Medical University, School of Medicine, Nagakute-cho, Aichi-gun, Aichi Prefecture, iologic Society of Japan14 and with the previous approval Japan. Address electronic mail to: [email protected]. Individual article reprints may be purchased through the Journal Web site, www.anesthesiology.org. of the Animal Care Committee of Aichi Medical Univer-

Anesthesiology, V 102, No 6, Jun 2005 1182 ISOFLURANE AND SEVOFLURANE EFFECTS ON NPE 1183 sity (Aichi-gun, Aichi Prefecture, Japan). Wistar male rats side of the cranium using a needle (26 gauge, 10 mm weighing 180–200 g (8–10 weeks old) were used. long). The vagus nerves were left intact. Rats were con- secutively treated with intracisternal injections of fibrin- Protocol ogen and thrombin, 0.075 ml each, at concentrations of The rats were randomly divided into three groups: 100 mg/ml and 200 U/ml, respectively. The severity of group 1, exposed to room air as the control (28 animals); pulmonary edema was graded from 0 to 3, with 0 for group 2, exposed to 1.5% isoflurane (1.09 minimum none and 3 for severe.15 When edema fluid appeared in alveolar concentration [MAC]) plus room air for4h(26 the tracheal tubes within 10 min, it was collected in animals); and group 3, exposed to 2.5% sevoflurane plastic tubes for later analysis, and the grade of edema (1.05 MAC) plus room air for 4 h (26 animals). Rats were formation was classified as grade 3. When edema fluid

placed in transparent plastic boxes (depth 50 cm ϫ did not appear in the tracheal tubes within 10 min, the Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 height 50 cm ϫ width 80 cm) into which anesthetic chest walls were opened. Sometimes edema fluid then gases were pumped at the desired partial pressure spontaneously appeared in the tracheal tubes, and this through inlet tubes and an outdoor outlet tube. The was classified as grade 2. However, in a few cases, edema concentrations of the anesthetic agents in the boxes fluid only appeared in the tracheal tubes when the lungs were monitored along with the gases in the air using an were gently compressed, and this was classified as grade isoflurane, sevoflurane, and respiratory oxygen/carbon 1. When edema fluid did not appear even with such dioxide analyzer (Capnomac; Datex Instrumentarium compression, the grade was 0. Finally, in all rats, lungs Co. Ltd., Helsinki, Finland). In 12 animals in groups 2 and were dissected out, weighed, and dried at 80°C over- 3, 0.1 ml anti-VEGF antibody (rabbit polyclonal anti- night. The difference between wet and dry weights, mouse VEGF immunoglobulin G; VEGF Ab-1, 1 mg/ml; relative to the dried lung weight, was used as the lung NeoMarkers, Fremont, CA) was injected into the exter- water ratio, an index of the severity of pulmonary nal cervical vein before the exposure to 1.5% isoflurane edema. or 2.5% sevoflurane. After this 4-h period, animals were moved outside of Immunohistochemistry VEGF mRNA Analysis the plastic boxes used for exposure to inhaled anesthet- Immunohistochemical and VEGF messenger RNA ics and then anesthetized with intraperitoneal injections (mRNA) expression studies were performed in two rats of pentobarbital sodium (35 mg/kg body weight) for the from groups 2 and 3 and four rats from group 1 (con- induction of NPE (see next two paragraphs). Tracheal trols). After opening the chest by a central incision with tubes were inserted after a tracheotomy was performed the aid of mechanical ventilation, a plastic catheter for in the midcervical region. Catheters were introduced infusion was inserted into the pulmonary artery through into the right femoral vein and artery, the former for the right ventricle. Blood was washed out with saline, blood sampling, and the latter for measurement of arte- and then paraform aldehyde, dissolved with phosphate rial blood pressure and heart rates (Multipurpose Poly- buffer (pH 7.4) at a concentration of 4%, was infused at gram, W-1100; Nihon Kohden, Tokyo, Japan). To measure a constant pressure of 80 cm H2O. Thereafter, one lung arterial oxygen tension immediately before the induction of from each animal was excised, immersed in the same NPE, 0.1 ml arterial blood was collected from rat femoral fixative for 12 h, and embedded in paraffin with an arteries. Arterial oxygen tension was measured with a auto–paraffin-embedding apparatus (Thermolyne Histo- blood gas analyzer (i-STAT and cartridge EG6ϩ; Abbott Center; Shiraimatsu Co. Ltd., Osaka, Japan). Laboratories Ltd., South Pasadena, CA). Paraffin blocks were made and sections were cut and In an additional 18 rats, after4hofexposure to 1.5% subjected to deparaffinization, hydration, and incubation isoflurane, NPE was induced during continued isoflurane with blocking serum (3% normal goat serum; Vector administration, rather than during pentobarbital anesthe- Laboratories, Burlingame, CA) before exposure to the sia. Nine of these animals were pretreated with intrave- primary antibody or preimmune serum. For negative nous injections of anti-VEGF antibody. In these animals, controls, preimmune rabbit immunoglobulin G was used an endotracheal tube was inserted during the anesthesia at the same concentration. After2hofincubation at through a mask (plastic cylinder; 5 cm diameter and room temperature, the slides were washed in phosphate- 15 cm length), which was connected to the outlet tube buffered saline and incubated for 30 min with a biotin- from the plastic gas-boxes mentioned above. Thereafter, ylated goat anti-rabbit antibody diluted 1:300 in phos- isoflurane was continuously introduced through the en- phate-buffered saline. We then used the avidin-biotin- dotracheal tube connected to the mask while NPE was complex detection method (Vectastain Elite ABC kit; induced. Vector Laboratories), with DAB as the substrate (DAKO Neurogenic pulmonary edema was induced as previ- Liquid DAB substrate-chromogen System; DAKO Corp., ously described.15 Briefly, the animals were fixed in a Carpinteria, CA), to detect VEGF localization. Sections prone position with a stereotaxic instrument, and the were subsequently dehydrated, mounted, and analyzed cisterna magna was accessed at the base of the dorsal by observers blinded to treatment assignment.

Anesthesiology, V 102, No 6, Jun 2005 1184 KANDATSU ET AL.

Lung homogenates and cultured rat aortic endothelial cells (see next section) exposed to isoflurane or sevoflu- rane were used for mRNA analyses to estimate VEGF expression. Reverse transcription polymerase chain re- action was performed with commercially provided re- agents on Ready-To-Go reverse transcription polymerase chain reaction beads (Amersham Pharmacia Biotech, Inc., Piscataway, NJ). The program for the thermal cycler (TaKaRa PCR Thermal Cycler MP; Takara Co. Ltd., To- kyo, Japan) was as follows: 95°C for 1 min, 50°C for 1 ϩ min, 72°C for 2 min, for 30 cycles, with 0.1 ␮g poly(A) Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 mRNA as the template and pd(T)12–18 as the first-strand primer. The polymerase chain reaction primers for VEGF Fig. 1. Typical results for systemic arterial pressure and heart were as follows: sense, 5'-TTTCGTAATTGAGGACATGG- rate. The downward arrow indicates edema fluid appearance in -consecutive injec ؍ antisense, 5'-AGAAAAGCCTATGATCTGAG-3'. The the tracheal tube at this time point. F ؉ T ;'3 ؍ tions of fibrinogen and thrombin into the fourth ventricle; PB polymerase chain reaction products, analyzed by elec- intraperitoneal injection of pentobarbital sodium. trophoresis on agarose gels, were normalized for expres- ␤ sion of the -actin housekeeping gene as an internal analyzer (Capnomac). Exposure was for 4 h before sam- control: The polymerase chain reaction products ob- pling of cells for reverse transcription polymerase chain ␤ tained for VEGF and -actin were 448 and 764 base pairs, reaction. respectively. Statistics Isolation and Culture of Rat Aortic Endothelial All results are expressed as mean Ϯ SD, with n denot- Cells ing the number of preparations (the number of animals). Rat aortic endothelial cells were isolated from Wistar Parametric data such as blood pressure, heart rate, and rats (150–200 g, 7–9 weeks old, male) and cultured lung water ratio were analyzed using paired t tests or according to the methods of Suh et al.16 Briefly, rats analyses of variance performed with Scheffé multiple were anesthetized with intraperitoneal injections of pen- comparison tests. When the SDs obtained in the para- tobarbital sodium, 35 mg/kg body weight, and their metric tests were significantly different among the aortas were removed and placed in phosphate-buffered ϩ ϩ groups, a nonparametric statistical method, the Kruskal- saline without Ca2 or Mg2 . The vessels were cleaned, Wallis one-way analysis of variance with Dunnett tests, opened longitudinally, cut into two or three small was performed. For nonparametric data such as inci- pieces, and placed with their intimal side down on dence, the chi-square test was used. P Ͻ 0.05 was con- Matrigel-coated plates in the growth medium. The sidered significant. growth medium contained 10% FCS, 75 ␮g/ml ECGS, 10 U/ml heparin, 100 U/ml penicillin–streptomycin, 1% L- glutamine, and 100 ␮M MEM nonessential amino acids, in Results Dulbecco’s modified eagle medium. After 4–7 days, the pieces were removed and cells harvested (more than Hemodynamics 90% viability of primary cultured cells was obtained Both systemic blood pressure and heart rate in the when assessed with trypan blue). The cells were identi- isoflurane and sevoflurane groups were lower than in fied as endothelial cells when they exhibited positive the rats maintained in room air (control group) (fig. 1 binding of anti–von Willebrand factor antibody (Dako- and table 1). Consecutive injections of fibrinogen and Cytomation, Carpinteria, CA), after fixation with 7:3 thrombin into the fourth ventricle greatly increased sys- methanol and acetone. temic arterial pressure and heart rate (fig. 1). The peak Culture plates containing rat aortic endothelial cells values for average systemic arterial pressure and heart were moved to another carbon dioxide incubator to rate obtained in the isoflurane and sevoflurane groups allow incubation under a desired partial pressure (1.5% were significantly lower than those obtained in the con- for isoflurane and 2.5% for sevoflurane) of inhaled anes- trol group (both P Ͻ 0.001). Changes in systemic arterial thetics at 37°C. The incubator was connected to a vinyl pressure in the isoflurane group but not in the sevoflu- bag in which 5% CO2, 20% O2, and desired amounts of rane group were significantly lower than those observed nitrogen gas and inhaled anesthetics had been mixed in the control group (P Ͻ 0.001). Changes in heart rate and pumped at an appropriate velocity. At the outlet, in the isoflurane group, in contrast, were significantly partial pressures of carbon dioxide and oxygen were greater than those obtained in the controls (P Ͻ 0.05). monitored along with inhaled anesthetics (isoflurane and Changes in heart rate in the sevoflurane group were sevoflurane) using a respiratory oxygen/carbon dioxide significantly smaller than with isoflurane (P Ͻ 0.05).

Anesthesiology, V 102, No 6, Jun 2005 ISOFLURANE AND SEVOFLURANE EFFECTS ON NPE 1185

Table 1. Mean Systemic Arterial Pressure and Heart Rate before and after Intrathecal Injections of Fibrinogen and Thrombin

Before Fibrin After Fibrin Changes in During Fibrin Pretreatment Treatment MSAP, mmHg HR, beats/min MSAP, mmHg HR, beats/min MSAP, mmHg HR, beats/min

Room air Pentobarbital 132.1 Ϯ 4.5 391.6 Ϯ 30.7 190.0 Ϯ 14.3 431.3 Ϯ 26.4 57.9 Ϯ 11.7 39.7 Ϯ 10.7 Antibody ϩ Pentobarbital 115.7 Ϯ 5.7§ 403.7 Ϯ 15.2 169.3 Ϯ 8.5§ 458.5 Ϯ 30.0 53.7 Ϯ 11.8 54.8 Ϯ 26.8‡ room air Isoflurane Pentobarbital 105.8 Ϯ 11.3† 227.2 Ϯ 43.6† 146.3 Ϯ 13.8† 281.9 Ϯ 40.1† 40.4 Ϯ 9.6† 54.8 Ϯ 26.4* Antibody ϩ Pentobarbital 84.1 Ϯ 5.4§ 303.3 Ϯ 48.3§ 165.3 Ϯ 7.7§ 389.2 Ϯ 15.1§ 81.2 Ϯ 5.3§ 85.8 Ϯ 37.3‡ isoflurane Sevoflurane Pentobarbital 84.3 Ϯ 6.5†# 232.8 Ϯ 28.9† 145.0 Ϯ 9.0† 278.3 Ϯ 36.3† 60.8 Ϯ 4.6# 45.5 Ϯ 10.0࿣ Antibody ϩ Pentobarbital 85.9 Ϯ 7.4 244.7 Ϯ 37.4 151.9 Ϯ 7.2 288.2 Ϯ 35.1 66.0 Ϯ 5.3 43.5 Ϯ 13.0 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 sevoflurane

Statistical significance at * P Ͻ 0.05 and † P Ͻ 0.001 compared with room air without rabbit anti–vascular endothelial growth factor antibody. Statistical significance at ‡ P Ͻ 0.05 and § P Ͻ 0.001 compared to without antibody. Statistical significance at ࿣ P Ͻ 0.05 and # P Ͻ 0.001 compared with isoflurane. Changes in mean systemic arterial pressure (MSAP) and heart rate (HR) caused by injections of fibrinogen and thrombin (fibrin) were all statistically significant (P Ͻ 0.001, with paired t test). When the SDs differed significantly between groups, a nonparametric statistical method, Kruskal-Wallis one-way analysis of variance with Dunnett tests, was used. Arterial oxygen tension after4hofinhalation with Effects of Anti-VEGF Antibody on NPE Development anesthetic gases (n ϭ 9, each group) were 87 Ϯ 6 and Mean systemic arterial pressure obtained in the isoflu- 86 Ϯ 6 mmHg for isoflurane and sevoflurane, respec- rane group after exposure to the anti-VEGF antibody was tively, these values being similar to those measured in significantly lower and the peak pressure after the injec- rats anesthetized with pentobarbital alone (85 Ϯ 3 tions of fibrinogen and thrombin was significantly higher mmHg). than those obtained without the pretreatment (both P Ͻ 0.001; table 1). The heart rate in the isoflurane group, Effects of Isoflurane and Sevoflurane on NPE when compared with that without the antibody, was Development significantly greater (P Ͻ 0.001), as was that after fibrin In the control group, 7 of 12 rats showed grade 2 or 3 treatment (P Ͻ 0.001). Changes in blood pressure and fibrin-induced pulmonary edema, with 5 exhibiting heart rate were significantly greater than without the grade 0 (table 2). Exposure to 1.5% isoflurane for 4 h Ͻ Ͻ resulted in a 100% incidence of grade 3 pulmonary antibody (P 0.001 and P 0.05, respectively). In the edema, significantly higher than the incidence in the sevoflurane and room air groups, no significant differ- control group (P Ͻ 0.05). After exposure to 2.5% ences in either systemic arterial pressure or heart rate sevoflurane for 4 h, one showed grade 3, and 11 showed were observed between the cases with and without grade 1 or 0, the 8% incidence of grade 2–3 pulmonary anti-VEGF antibody pretreatment. edema being significantly lower than that in the isoflu- Pretreatment with antibody did not significantly affect rane group (P Ͻ 0.001). arterial blood oxygen tension obtained after4hofex- The lung water ratio, an index of pulmonary edema posure to isoflurane (89 Ϯ 4 mmHg, n ϭ 9). severity, was significantly higher in the isoflurane group All 12 rats exposed to 1.5% isoflurane for 4 h were compared with controls (table 2). In contrast, the lung successfully protected from fibrin-induced pulmonary water ratio did not differ between control and sevoflu- edema development by pretreatment with the anti-VEGF rane-anesthetized rats (table 2). antibody, the incidence being lower than that obtained Table 2. Incidence (Grades 2 and 3) and Severity of Fibrin-induced Pulmonary Edema

Grade of Edema Positive Rate 0123 (Grades 2 and 3), % Lung Water Ratio

Room air ϩ pentobarbital 5 0 1 6 58 4.86 Ϯ 0.78 Anti-VEGF antibody ϩ room air ϩ 4 0 1 7 67 5.04 Ϯ 0.45 pentobarbital 1.5% Isoflurane ϩ pentobarbital 00012 100* 6.15 Ϯ 0.64† Anti-VEGF antibody ϩ 1.5% 12 0 0 0 0‡ 3.74 Ϯ 0.21‡ isoflurane ϩ pentobarbital 2.5% Sevoflurane ϩ pentobarbital 10 1 0 1 8§ 4.40 Ϯ 0.32§ Anti-VEGF antibody ϩ 2.5% 11 1 0 0 0 4.22 Ϯ 0.25 sevoflurane ϩ pentobarbital

Statistical significance at * P Ͻ 0.05 and † P Ͻ 0.001 compared with room air ϩ pentobarbital without anti–vascular endothelial growth factor (VEGF) antibody. ‡ Statistical significance at P Ͻ 0.001 compared to without anti-VEGF antibody. § Statistical significance at P Ͻ 0.001 compared with isoflurane. Statistical significance for incidence was analyzed with the chi-square test.

Anesthesiology, V 102, No 6, Jun 2005 1186 KANDATSU ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021

Fig. 3. Immunohistochemical staining of vascular endothelial growth factor in lungs. (A) Stained with unimmunized rabbit serum as the primary antibody. (B) Lungs of a rat anesthetized with 2.5% sevoflurane for 4 h. (C–F) Lungs of rats anesthetized with 1.5% isoflurane for 4 h. Horizontal bars below the figures Fig. 2. Vascular endothelial growth factor messenger RNA indicate 100 ␮m. (mRNA) expression in lung homogenates and primary mono- layer cultures of aortic endothelial cells. Reverse transcription Immunohistochemistry for VEGF Staining polymerase chain reaction products were analyzed by electro- Figure 3A shows staining with unimmunized rabbit phoresis (upper panel). Equal amounts of complementary DNA serum as the primary antibody. Immunohistochemical were amplified for vascular endothelial growth factor mRNA and ␤-actin mRNA at 30 cycles. Lanes 1, 2, and 3 are lung staining showed no apparent VEGF expression in the homogenates for sevoflurane, isoflurane, and room air groups, lungs of the sevoflurane group, as shown in figure 3B, respectively. Lanes 4, 5, and 6 are aortic endothelial monolay- whereas after isoflurane treatment (figs. 3C–F), bron- ers for the same groups. ␤-Actin was used as an internal stan- dard to confirm the mRNA integrity of each sample. Sizes for chial epithelial cells (figs. 3C–E), and smooth muscles vascular endothelial growth factor and ␤-actin are 448 and 764 (fig. 3D) were strongly immunoreactive. Vascular endo- base pairs (bp), respectively. The lower panel indicates the thelial cells were also positively stained, especially in the mean and SD of the vascular endothelial growth factor mRNA amount obtained from electrophoresis with computer analyses alveolar capillaries (fig. 3F). (OD multiplied by length). There were four experiments for each group. *** Statistical significance compared to without the Effects of Anti-VEGF Antibody on NPE Development antibody. in Rats Anesthetized with Isoflurane for NPE Induction Ͻ without the antibody (P 0.001; table 2). Furthermore, In the group with4hofisoflurane exposure with the lung water ratio in rats exposed to 1.5% isoflurane in isoflurane anesthesia during experiments for NPE induc- the presence of the anti-VEGF antibody was significantly tion, injections of fibrinogen and thrombin into the Ͻ lower than that obtained without the antibody (P fourth ventricle significantly increased the mean sys- 0.001). Rats maintained in room air or exposed to 2.5% temic arterial pressure from 97.8 Ϯ 12.3 to 144.6 Ϯ sevoflurane were not significantly influenced by the pre- 15.3 mmHg and significantly increased the mean heart treatment with the anti-VEGF antibody in terms of rate from 288.4 Ϯ 29.6 to 358.4 Ϯ 26.2 beats/min (both incidence. P Ͻ 0.001). The incidence of grade 2 and 3 fibrin- induced pulmonary edema and the lung water ratio were VEGF mRNA Expression in Lungs and Cultured 100% (nine grade 3 in nine animals), and 4.46 Ϯ 0.23, Aortic Endothelial Cells respectively. Intravenous injection of 0.1 ml anti-VEGF As shown in the upper panel of figure 2, the lungs in antibody before isoflurane exposure significantly de- isoflurane-exposed rats (lane 2) exhibited greater VEGF creased the incidence and lung water ratio to 0% (nine mRNA expression relative to ␤-actin mRNA compared grade 0 in nine animals) and 3.40 Ϯ 0.09 (fig. 4), with sevoflurane-exposed or control lungs (lanes 1 and respectively. 3, respectively). In contrast, aortic endothelial cells in- cubated in 2.5% sevoflurane plus 5% CO or 1.5% isoflu- 2 Discussion rane plus 5% CO2 (lanes 4 and 5) demonstrated similar expression compared with cells exposed to room air The results obtained in the current study showed the plus 5% CO2 (lane 6). development of fibrin-induced pulmonary edema to be

Anesthesiology, V 102, No 6, Jun 2005 ISOFLURANE AND SEVOFLURANE EFFECTS ON NPE 1187

Isoflurane is an inhaled anesthetic that primarily re- laxes the vasculature in the systemic cardiovascular sys- tem to exert hypotensive effects, in contrast to sevoflu- rane, which mainly diminishes cardiac output.18 Reports document increased alveolar epithelial and capillary per- meability in humans exposed to 1.5% isoflurane for ap- proximately 6 h.7,8,19 In contrast, sevoflurane elicited an inhibitory action on the capillary filtration coefficient in the lower limbs when assessed with noninvasive com- puter-assisted venous congestion.20 No significant differ-

ence between isoflurane and sevoflurane in terms of Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 pulmonary vascular permeability was found with short- duration volatile anesthetic exposure; 30 min of expo- sure to either isoflurane or sevoflurane before ischemia in isolated rat lungs attenuated the increase of vascular permeability with ischemia–reperfusion injury.21 There- fore, it is likely that long exposure to 1.5% isoflurane causes interstitial edema and thus enhances NPE devel- opment, unlike sevoflurane, but the exact mechanisms remain to be clarified in detail. Fig. 4. Lung water ratios obtained with intracisternal fibrin The current immunohistochemical study demon- injection during isoflurane anesthesia continually performed strated expression of VEGF to be increased in the lungs after4hofisoflurane exposure with or without rabbit anti– obtained from rats anesthetized with isoflurane when vascular endothelial growth factor antibody pretreatment. Fi- brinogen and thrombin were injected into the cisterna magna compared with sevoflurane, suggesting that the increase of rats during anesthesia with 1.5% isoflurane via tracheal in lung vascular permeability may be caused via up- tubes. Rabbit anti–vascular endothelial growth factor antibody regulation of VEGF, a potent inducer of increased per- (Antibody) was intravenously injected into the femoral vein (0.1 ml) before the4hofexposure to 1.5% isoflurane. meability. As shown in figure 3, an increase in VEGF *** Statistical significance at a level of 0.001, compared to with- expression was found in bronchial epithelial and smooth out the antibody. muscle cells, as well as alveolar capillary endothelial cells. Nonetheless, no marked change was observed in promoted by isoflurane and inhibited by sevoflurane. primary cultured endothelial monolayers. This may sug- Because the systemic arterial blood pressure after 4 h of gest that bronchial epithelial cells or other cells found in exposure to the two agents did not differ significantly, the lung homogenate, rather than endothelial cells, may the difference in incidence cannot be explained simply initially respond to isoflurane exposure. Alternatively, by this parameter. The inhibitory effects of sevoflurane the cultured monolayers may not reflect the situation on NPE development may correspond well with the found in vivo. A wide variety of cell types may express 17 decrease in systemic arterial pressure, in accord with VEGF and VEGF receptors, and therefore, activated mac- the facts that inhaled anesthetics usually diminish ner- rophages, neutrophils, and alveolar type II epithelial vous activity and that NPE development is associated cells cannot be ruled out as candidate targets.22 It is 3,11 with enhanced sympathetic nerve activity. However, possible that VEGF released from the lungs may affect contrary to expectations, the incidence of fibrin-induced both the alveolar epithelial and endothelial barriers, pulmonary edema was increased by exposure to thereby resulting in gas exchange dysfunction. isoflurane. Intravenous injection of an anti-VEGF antibody before We examined the effects of isoflurane using two ex- exposure of rats to 1.5% isoflurane was found to inhibit perimental designs. In the first, animals were pretreated NPE development completely, whereas in rats anesthe- with4hofisoflurane and then anesthetized with pen- tized with pentobarbital or sevoflurane, pretreatment tobarbital for the induction of NPE. It is possible that this with the antibody did not affect the incidence of NPE. experimental design, which involves two different anes- These results suggested that VEGF plays a causative role thetics in a similar experimental animal, could compli- in the enhancement of NPE development by isoflurane. cate the interpretation of the results. For this reason, we It has been reported4 that both VEGF and neuropeptide performed an additional set of studies in which isoflu- Y, a coneurotransmitter with norepinephrine in the sym- rane was maintained during the induction of NPE to pathetic nerve terminals, enhance albumin permeability eliminate any possible confounding effects of pentobar- across rat aortic endothelial monolayers and that such an bital. These results were qualitatively similar, with the action is stronger under hypoxic (5% O2) than normoxic exception that the absolute values of lung water ratio conditions. Neuropeptide Y has little effect on perme- were less in the latter protocol. ability under normoxic conditions but a potent action

Anesthesiology, V 102, No 6, Jun 2005 1188 KANDATSU ET AL. when the environment is hypoxic. It is possible that the respiratory distress syndrome,32 where the involvement release of VEGF during exposure to isoflurane may cause of neuropeptide Y and VEGF remains unclear. an interstitial edema in lungs that produces local hyp- oxia, which may synergistically enhance or cause a fur- ther increase in vascular permeability via release of neuropeptide Y, with its vasoconstrictive action and References action on sympathetic nerves. Pretreatment with anti- 1. Malik AB: Mechanisms of neurogenic pulmonary edema. Circ Res 1985; VEGF antibodies could prevent local generation of hy- 57:1–18 2. Hirabayashi A, Nishiwaki K, Shimada Y, Ishikawa N: Role of neuropeptide poxic conditions. However, systemic hypoxia was not Y and its receptor subtypes in neurogenic pulmonary edema. Eur J Pharmacol observed after isoflurane. Whatever the case, it is likely 1996; 296:297–305 3. Hamdy O, Maekawa H, Shimada Y, Feng GG, Ishikawa N: Role of central that the increase in VEGF expression observed with nervous system nitric oxide in the development of neurogenic pulmonary edema Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021 isoflurane exposure makes an essential contribution to in rats. Crit Care Med 2001; 29:1222–8 4. Nan YS, Feng GG, Hotta Y, Nishiwaki K, Shimada Y, Ishikawa A, Kurimoto enhanced NPE development. N, Shigei T, Ishikawa N: Neuropeptide Y enhances permeability across a rat Isoflurane is a commonly used volatile anesthetic agent aortic endothelial cell monolayer. Am J Physiol Heart Circ Physiol 2004; 286: 23 H1027–33 because of its low toxicity. It is known to increase the 5. Donoso MV, Miranda R, Irarrazaval MJ, Huidobro-Toro JP: Neuropeptide Y 24 nitric oxide content in the rat brain cortex, an effect is released from human mammary and radial vascular biopsies and is a functional modulator of sympathetic cotransmission. J Vasc Res 2004; 41:387–99 abolished by nitric oxide synthase inhibition. Isoflurane 6. Featherstone RM, Muehlbaecher CA: The current role of inert gases in the increases intracellular calcium concentrations in cere- search for anesthesia mechanisms. Pharmacol Rev 1963; 15:97–121 brocortical neurons.25 Therefore, isoflurane may in- 7. Sun SS, Hsieh JF, Tsai SC, Ho YJ, Kao CH: Transient increase in alveolar epithelial permeability induced by volatile anesthesia with isoflurane. Lung 2000; crease nitric oxide concentrations in the cortex, which 178:129–35 8. Changlai SP, Hung WT, Liao KK: Detecting alveolar epithelial injury follow- may diminish autonomic nerve activity, especially sym- 99m 26 ing volatile anesthetics by TcDTPA radioaerosol inhalation lung scan. Respi- pathetic nerve activity. It could also affect peripheral ration 1999; 66:506–10 nitric oxide concentrations in the tissue, which are also 9. Shayevitz JR, Johnson KJ, Knight PR: Halothane–oxidant interactions in the ex vivo perfused rabbit lung: Fluid conductance and eicosanoid production. important regulators of endothelial functions affecting ANESTHESIOLOGY 1993; 79:129–38 vascular tone, endothelial cell permeability, and vascular 10. Rezaiguia-Delclaux S, Tayr C, Luo DF, Saïdi NE, Meignan M, Duvaldestin P: 27 3,26 Halothane and isoflurane decrease alveolar epithelial fluid clearance in rats. cell proliferation. Two studies have shown that an ANESTHESIOLOGY 1998; 88:751–60 increase in nitric oxide content in the nucleus tractus 11. Sakakibara H, Hashiba Y, Taki K, Kawanishi M, Shimada Y, Ishikawa N: Effects of sympathetic nerve stimulation on lung vascular permeability in the rat. solitarii was associated with a decrease in the incidence Am Rev Resp Dis 1992; 145:685–92 of NPE. However, sevoflurane is also reported to activate 12. Dvorak HF, Detmar M, Claffey KP, Nagy JA, van de Water L, Senger DR: nitric oxide synthase to increase central nitric oxide Vascular permeability factor/vascular endothelial growth factor: An important mediator of angiogenesis in malignancy and inflammation. Int Arch Allergy 24 concentrations in the rat brain cortex, making it less Immunol 1995; 107:233–5 likely that this mechanism contributes to the observed 13. Barleon B, Siemeister G, Martiny-Baron G, Weindel K, Herzog C, Marme D: Vascular endothelial growth factor up-regulates its receptor fms-like tyrosine effects of isoflurane. kinase 1 (FLT-1) and a soluble variant of FLT-1 in human vascular endothelial Vascular endothelial growth factor exerts a direct in- cells. Cancer Res 1997; 57:5421–5 14. Physiological Society of Japan: Guiding principles for the care and use of fluence on endothelial cells through VEGFR-2–coupled animals in the field of physiological sciences [in Japanese]. Nippon Seirigaku tyrosine kinase, activating protein kinase C28 and in- Zasshi 2002; 64:143–6 15. Ishikawa N, Kainuma M, Furuta T, Sato Y: Factors influencing fibrin- creasing the activity of endothelial nitric oxide syn- induced pulmonary edema. Jpn J Pharmacol 1988; 46:255–60 ϩ thase.27 It has been shown to increase the Ca2 influx 16. Suh SH, Vennekens R, Manolopoulos VG, Freichel M, Schweig U, Prenen J, Flockerzi V, Droogmans G, Nilius B: Characterisation of explanted endothelial across the plasma membrane in endothelial cells in cells from mouse aorta: Electrophysiology and Ca2ϩ signaling. Pflugers Arch vitro,29,30 and nitric oxide synthesized and released in 1999; 438:612–20 17. Nishiwaki K, Hirabayashi A, Shimada Y, Ishikawa N: Effects of vasodilators response to VEGF may also contribute to increases of on fibrin-induced pulmonary edema, so-called neurogenic pulmonary edema, in 31 vascular permeability. the rat. J Anesth 1994; 8:208–12 In conclusion, exposure to isoflurane for more than 18. Stoelting RK: Inhaled anesthetics, Pharmacology and Physiology in Anesthetic Practice, 2nd edition. Edited by Stoelting RK. Lippincott, 1991, pp 4 h may enhance vascular permeability across lung vas- 33–69 cular endothelial cells via release of VEGF. This may 19. Hung CJ, Liu FY, Wu RS, Tsai JJ, Lin CC, Kao A.: The influence of volatile anesthetics on alveolar epithelial permeability measured by noninvasive radionu- cause interstitial edema, resulting in local hypoxia, clide lung scan. Ann Nucl Med 2003; 17:213–8 which then enhances NPE development. In contrast, 20. Bruegger D, Bauer A, Finsterer U, Bernasconi P, Kreimeier U, Christ F: Microvascular changes during anesthesia: Sevoflurane compared with propofol. exposure to sevoflurane is not associated with such Acta Anaesthesiol Scand 2002; 46:481–7 effects. Because NPE is linked to overexcitation of sym- 21. Liu R, Ishibe Y, Ueda M: Isoflurane-sevoflurane administration before ischemia attenuates ischemia-reperfusion–induced injury in isolated rat lungs. pathetic nerves, neuropeptide Y, a coneurotransmitter ANESTHESIOLOGY 2000; 92:833–40 with norepinephrine, may participate in its enhanced 22. Webb NJA, Myers CR, Watson CJ, Bottomley MJ, Brenchley PEC: Activated human neutrophils express vascular endothelial growth factor (VEGF). Cytokine development, especially under conditions where VEGF 1998; 10:254–7 expression is increased. It is still unclear whether anes- 23. Elliott RH, Strunin L: Hepatotoxicity of volatile anesthetics. Br J Anaesth 1993; 70:339–48 thesia with isoflurane is closely related to pulmonary 24. Baumane L, Dzintare M, Zvejniece L, Meirena D, Lauberte L, Sile V, edema induced by surgical maneuvers, such as the acute Kalvinsh I, Sjakste N: Increased synthesis of nitric oxide in rat brain cortex due

Anesthesiology, V 102, No 6, Jun 2005 ISOFLURANE AND SEVOFLURANE EFFECTS ON NPE 1189

to halogenated volatile anesthetics confirmed by EPR spectroscopy. Acta Anaes- permeability across rat endothelial monolayers, induced b neuropeptide Y or thesiol Scand 2002; 46:378–83 VEGF. Am J Physiol Heart Circ Physiol 2004; 287:H100–6 25. Kindler CH, Eilers H, Donohoe P, Ozer S, Bickler PE: Volatile anesthetics 29. Pocock TM, Williams B, Curry FE, Bates DO: VEGF and ATP act by different 2ϩ increase intracellular calcium in cerebrocortical and hippocampal neurons. AN- mechanisms to increase microvascular permeability and endothelial [Ca ]i. ESTHESIOLOGY 1999; 90:1137–45 Am J Physiol Heart Circ Physiol 2000; 279:H1625–34 26. Feng GG, Nishiwaki K, KondoH, Shimada Y, Ishikawa N: Inhibition of 30. Bates DO, Curry FE: Vascular endothelial growth factor increases micro- fibrin-induced neurogenic pulmonary edema by previous unilateral left vagotomy vascular permeability via a Ca(2ϩ)-dependent pathway. Am J Physiol 1997; correlates with increased levels of brain nitric oxide synthase in the nucleus 273:H687–94 tractus solitarii. Auton Neurosci 2002; 102:1–7 31. Murohara T, Horowitz JR, Silver M, Tsurumi Y, Chen D, Sullivan A, Isner 27. Kroll J, Waltenberger J: A novel function of VEGF receptor-2 (KDR): Rapid JM: Vascular endothelial growth factor/vascular permeability factor enhances vascu- release of nitric oxide in response to VEGF-A stimulation in endothelial cells. lar permeability via nitric oxide and prostacyclin. Circulation 1998; 97:99–107 Biochem Biophys Res Commun 1999; 265:636–9 32. Thickett DR, Armstrong L, Christie SJ, Millar AB: Vascular endothelial 28. Kurimoto N, Nan YS, Chen ZY, Feng GG, Komatsu T, Kandatsu N, Ko J, growth factor may contribute to increased vascular permeability in acute respi- Kawai N, Ishikawa N: Effects of specific signal transductic inhibitors on increased ratory distress syndrome. Am J Respir Crit Care Med 2001; 164:1601–5 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/6/1182/359309/0000542-200506000-00018.pdf by guest on 30 September 2021

Anesthesiology, V 102, No 6, Jun 2005