Monoclonal
Antibody to Tumor Necrosis Factor \g=a\ Attenuates Cardiopulmonary Dysfunction in Porcine Gram-negative Sepsis
Ciaran J. Walsh, FRCSI; Harvey J. Sugerman, MD; Patrick G. Mullen, FRCSI; P. Declan Carey, FRCSI; Sandra K. Leeper-Woodford, PhD; Gary J. Jesmok, PhD; Earl F. Ellis, PhD; Alpha A. Fowler, MD
necrosis factor (TNF) is implicated in the pathophysiology of gram-negative sepsis. This study examined physiologic and biochemical effects of pretreatment with an anti-TNF\g=a\monoclonal antibody immediately before the onset of sepsis. Three groups of anesthetized ventilated pigs were studied for 300 minutes. Groups 1 (n=12) and 2 (n=6) received a 1-hour infusion of live Pseudomonas aeruginosa. Group 2 was pretreated with anti-TNF\g=a\monoclonal antibody (15 mg/kg). Group 3 (n=8) received intravenous sterile saline. Group 1 exhibited a significant rise in plasma TNF activity, which was abolished in group 2. Cardiac index was reduced in both groups 1 and 2 in the first hour but recovered in group 2 (3.3\m=+-\0.4L/min per square meter at 300 minutes in group 2 vs 1.3\m=+-\0.2 L/min per square meter in group 1). Metabolic acidosis was attenuated (arterial pH, 7.39\m=+-\0.01in group 2 vs 7.16\m=+-\0.03at 300 minutes in group 1). Increased extravascular lung water was also attenuated (5.9\m=+-\0.7in group 2 vs 13.2\m=+-\1.5mL/kg at 300 minutes in group 1). However, pulmonary hypertension and hypoxemia, which are known cyclooxygenase effects, were not affected. In the early phase of the study, plasma thromboxane B2 levels were elevated in both groups 1 and 2. We conclude that anti-TNF\g=a\monoclonal antibody offered significant protection against the effects of sepsis, but that other mediators may be responsible for the early changes seen in this model. \s=b\ Tumor
(Arch Surg. 1992;127:138-145) advances antibiotic therapy and the Despite treatment, septic shock of death intensive in
intensive is most common ' cause care units. Sepsis, particularly in abdominal sepsis, remains the most common predisposicare
Accepted
for
publication
October 20, 1991.
Departments of Surgery (Drs Walsh, Sugerman, Mullen, and Carey), Internal Medicine (Drs Leeper-Woodford and Fowler), Pharmacology (Dr Ellis), and Pathology (Drs Leeper\x=req-\ Woodford and Fowler), Medical College of Virginia, Virginia Commonwealth University, Richmond; and Cutter Biological, Miles Inc, Berkeley, Calif (Dr Jesmok). Presented at the 11th Annual Meeting of the Surgical Infection Society, Fort Lauderdale, Fla, April 8, 1991. Reprint requests to PO Box 50, MCV Station, Medical College of Virginia, Richmond, VA 23298-0519 (Dr Fowler). From the
tion for the adult respiratory distress syndrome.2-3 When adult respiratory distress syndrome is associated with sepsis, mortality approaches 80%.4 Improvement in pa¬ tient outcome will on early recognition of patients at high risk of developing sepsis and multiple organ fail¬ ure as well as the development of effective early pharma¬ cologie intervention aimed at specific mediators of the
depend
septic process.
It has become increasingly clear in recent years that the physiological and biochemical effects of bacteria and bac¬ terial endotoxin are largely mediated through a network of cytokines interacting with cyclooxygenase and lipoxygenase metabolites of arachidonic acid.2 Tumor necrosis
factor
(TNF), a proinflammatory cytokine elaborated pre¬
dominantly by cells
of macrophage/monocyte lineage in response to bacterial endotoxin,5 is now recognized as a primary mediator of the host response to gram-negative sepsis.6-8 Pretreating mice,9 rabbits,10 and baboons1112 with anti-TNF monoclonal antibody (MoAb) reduces animal mortality after administration of endotoxin or lethal numbers of gram-negative organisms. The mechanisms acting to reduce mortality after pretreatment with antiTNF MoAb have not been fully elucidated. The present study examines the effects of anti-TNFa MoAb pretreatment on the course of gram-negative sep¬ sis and acute lung injury produced by infusion of live Pseudomonas aeruginosa in the pig. The pretreatment design of this study permits the pathogenic effects of TNF to be examined. Mounting evidence suggests that many of the effects of TNF occur via cyclooxygenase metabo¬ lites.7 We set out to examine these relationships between TNF and cyclooxygenase metabolites in porcine sepsis by studying the effects of anti-TNFa MoAB pretreatment on plasma thromboxane levels. CLINICAL SIGNIFICANCE Clinical trials are presently in progress to evaluate antiTNFa MoAb as a therapeutic option in gram-negative sepsis.13 However, the effects of blocking TNFa are not fully elucidated and there is concern over potential dele¬ terious effects of blocking the normal cytokine response.8 The early TNF surge in plasma and subsequent cytokine cascade associated with cyclooxygenase and lipoxygenase
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Michigan User on 06/16/2015
metabolites of arachadonic acid represent a complex bio¬ logical response to bacterial invasion. Gathering informa¬ tion concerning the physiological and biochemical effects of TNFa blockade is central to a better understanding of the biological effects of this therapeutic option. MATERIALS AND METHODS Anti-TNFa MoAb The IgGl anti-TNFa MoAB used was purified from murine hybridoma culture harvests via cell separation, polyethylene glycol precipitation, anion exchange, and size exclusion chromatography. Purified MoAb was 99% pure with fully functional binding to human TNF. Endotoxin levels were less than 2 pg/mg of protein (limulus assay). Stabilization was performed with glycine and maltose before lyophilization. Lyophilized MoAb was kept at 4°C until the time of the experiment, when it was re¬ constituted with 20 mL of sterile water. Anti-TNFa MoAb (15 mg/kg) was administered for 10 minutes via the left internal jug¬ ular vein.
Porcine Model Yorkshire pigs weighing between 15 and 33 kg were obtained from a commercial vendor and kept in the Virginia Common¬ wealth University vivarium for 3 to 5 days before use. All animals received intramuscular penicillin G benzethine and penicillin G procaine (300 000 U each) 48 hours before use as part of preop¬ erative conditioning. Animals were preanesthetized with intra¬ muscular ketamine hydrochloride (25 mg/kg) and placed supine. Pentobarbital sodium (20 mg/kg) was administered intrave¬ nously to induce anesthesia, which was maintained with con¬ tinuous pentobarbital infusion (1 mg/kg per hour). After ade¬ quacy of anesthesia was determined, animals were paralyzed with a continuous intravenous infusion of pancuronium bro¬ mide (0.2 mg/kg per minute). Trachéal intubation was performed with a cuffed endotracheal tube (National Catheter). Ventilation was performed with a large-animal ventilator (Harvard Appara¬ tus, Boston, Mass) employing a 50% fraction of inspired oxygen and 5 cm H20 positive end-expiratory pressure. Vascular cath¬ eters were placed in the left common carotid artery for monitor¬ ing systemic arterial pressure and determining arterial blood gas levels and in the left external jugular vein for infusion of bacte¬ ria. An indwelling balloon-tipped pulmonary arterial catheter was inserted via the right external jugular vein and positioned via pressure monitoring in the pulmonary artery for measure¬ ment of pulmonary arterial pressure and cardiac output (Ed¬ wards COMI, Santa Ana, Calif). Cardiac index (CI) was calcu¬ lated using the following formula:
Cardiac Output/[(Body Weight20) x 0.112] A 5F lung-water catheter (model 96-020-5F, American Edwards Laboratories, Tustin, Calif) was positioned in the lower abdom¬ inal aorta via the left femoral artery for measurement of extravascular lung water (EVLW) with the thermal indocyanine green (Cardio-Green) double indicator technique.14 Systemic ar¬ terial blood sample and mixed venous samples were drawn from the distal port of the pulmonary arterial catheter for blood gas analysis using a pH/blood gas analyzer (model 1304, Instrumen¬ tation Laboratories, Lexington, Mass). Hemodynamic and blood gas variables were measured at 30-minute intervals for the first hour and every hour thereafter. Extravascular lung water mea¬ surements were taken at baseline and at 60-minute intervals un¬ til the end of the 5-hour study period. CI
=
TNF
fuged at 500g for 20 minutes at 4°C and the resulting plasma was
frozen at 20°C until time of assay. The L929 cells were seeded into flat-bottomed 96-well microtiter plates (Corning Glass Works, Corning, NY) at a density of 4 x 104 cells per well and grown to confluence overnight in Dulbecco's minimal essential medium (DMEM; GIBCO BRL, Grand Island, NY) containing 1% penicillin-streptomycin and 5% fetal calf serum (DMEM). The medium was then removed from confluent monolayers and 100 µ of DMEM containing dactinomycin (Merck, Sharp and Dohme, Westpoint, Pa) in a final concentration of 5 mg/L was added to each well. One hundred microliters of each of the fol¬ lowing was then added to selected duplicate wells containing L929 cells: (1) DMEM (0% cytotoxicity); (2) serial dilutions of recombinant TNF (5xl0~3 to 6xl0~4 U/mL) (Cetus Corp, Em¬ eryville, Calif); (3) plasma samples from different groups; (4) DMEM in blank wells without cells (100% cytotoxicity). The plates were then incubated for 20 hours at 37°C in 5% carbon di¬ oxide. The medium was removed after incubation, and the L929 cells were stained for 10 minutes with 0.5% crystal violet in 20% methanol, rinsed in water, and air dried. Optical density of each well was determined with a microplate reader (Bio-Tek EL 309) and calibrated to noncellular reagent blanks at a wavelength of 550 nm. The percentage of cytotoxicity of L929 cells was calcu¬ lated with the following formula: % of Cytotoxicity = (Optical Density of Wells With 0% Cytotoxicity Optical Density of Experimental Sample Well)/ Optical Density of Wells With 0% Cytotoxicity Tumor necrosis factor activity is expressed in units per milliliter, with one unit of TNF activity defined as 50% L929 cytotoxicity. —
—
Plasma Thromboxane B2
—
Study Groups Three groups of animals were studied. Groups 1 (n 12) and 2 (n 6) received a 1-hour infusion of live aeruginosa (5xl08 colony-forming units [CFU]/mL at 0.3 mL/20 kg per minute). Group 2 was pretreated with 15 mg/kg of intravenous anti-TNFa MoAb 15 minutes before the Pseudomonas infusion. Group 3 (n 8) received sterile saline only. All animals were studied for 300 minutes. The experimental protocol for this study was approved by the Institutional Animal Care and Use Committee of Virginia Com¬ monwealth University and adhered to National Institutes of Health guidelines for the use of experimental animals. =
=
=
Statistical
Analysis
Results are expressed as means ± SEMs. Differences were tested for significance using analysis of variance. Differences between means were analyzed using Tukey's Studentized Range Test. Results were considered statistically significant at P
Antibody to Tumor Necrosis Factor \g=a\ Attenuates Cardiopulmonary Dysfunction in Porcine Gram-negative Sepsis
Ciaran J. Walsh, FRCSI; Harvey J. Sugerman, MD; Patrick G. Mullen, FRCSI; P. Declan Carey, FRCSI; Sandra K. Leeper-Woodford, PhD; Gary J. Jesmok, PhD; Earl F. Ellis, PhD; Alpha A. Fowler, MD
necrosis factor (TNF) is implicated in the pathophysiology of gram-negative sepsis. This study examined physiologic and biochemical effects of pretreatment with an anti-TNF\g=a\monoclonal antibody immediately before the onset of sepsis. Three groups of anesthetized ventilated pigs were studied for 300 minutes. Groups 1 (n=12) and 2 (n=6) received a 1-hour infusion of live Pseudomonas aeruginosa. Group 2 was pretreated with anti-TNF\g=a\monoclonal antibody (15 mg/kg). Group 3 (n=8) received intravenous sterile saline. Group 1 exhibited a significant rise in plasma TNF activity, which was abolished in group 2. Cardiac index was reduced in both groups 1 and 2 in the first hour but recovered in group 2 (3.3\m=+-\0.4L/min per square meter at 300 minutes in group 2 vs 1.3\m=+-\0.2 L/min per square meter in group 1). Metabolic acidosis was attenuated (arterial pH, 7.39\m=+-\0.01in group 2 vs 7.16\m=+-\0.03at 300 minutes in group 1). Increased extravascular lung water was also attenuated (5.9\m=+-\0.7in group 2 vs 13.2\m=+-\1.5mL/kg at 300 minutes in group 1). However, pulmonary hypertension and hypoxemia, which are known cyclooxygenase effects, were not affected. In the early phase of the study, plasma thromboxane B2 levels were elevated in both groups 1 and 2. We conclude that anti-TNF\g=a\monoclonal antibody offered significant protection against the effects of sepsis, but that other mediators may be responsible for the early changes seen in this model. \s=b\ Tumor
(Arch Surg. 1992;127:138-145) advances antibiotic therapy and the Despite treatment, septic shock of death intensive in
intensive is most common ' cause care units. Sepsis, particularly in abdominal sepsis, remains the most common predisposicare
Accepted
for
publication
October 20, 1991.
Departments of Surgery (Drs Walsh, Sugerman, Mullen, and Carey), Internal Medicine (Drs Leeper-Woodford and Fowler), Pharmacology (Dr Ellis), and Pathology (Drs Leeper\x=req-\ Woodford and Fowler), Medical College of Virginia, Virginia Commonwealth University, Richmond; and Cutter Biological, Miles Inc, Berkeley, Calif (Dr Jesmok). Presented at the 11th Annual Meeting of the Surgical Infection Society, Fort Lauderdale, Fla, April 8, 1991. Reprint requests to PO Box 50, MCV Station, Medical College of Virginia, Richmond, VA 23298-0519 (Dr Fowler). From the
tion for the adult respiratory distress syndrome.2-3 When adult respiratory distress syndrome is associated with sepsis, mortality approaches 80%.4 Improvement in pa¬ tient outcome will on early recognition of patients at high risk of developing sepsis and multiple organ fail¬ ure as well as the development of effective early pharma¬ cologie intervention aimed at specific mediators of the
depend
septic process.
It has become increasingly clear in recent years that the physiological and biochemical effects of bacteria and bac¬ terial endotoxin are largely mediated through a network of cytokines interacting with cyclooxygenase and lipoxygenase metabolites of arachidonic acid.2 Tumor necrosis
factor
(TNF), a proinflammatory cytokine elaborated pre¬
dominantly by cells
of macrophage/monocyte lineage in response to bacterial endotoxin,5 is now recognized as a primary mediator of the host response to gram-negative sepsis.6-8 Pretreating mice,9 rabbits,10 and baboons1112 with anti-TNF monoclonal antibody (MoAb) reduces animal mortality after administration of endotoxin or lethal numbers of gram-negative organisms. The mechanisms acting to reduce mortality after pretreatment with antiTNF MoAb have not been fully elucidated. The present study examines the effects of anti-TNFa MoAb pretreatment on the course of gram-negative sep¬ sis and acute lung injury produced by infusion of live Pseudomonas aeruginosa in the pig. The pretreatment design of this study permits the pathogenic effects of TNF to be examined. Mounting evidence suggests that many of the effects of TNF occur via cyclooxygenase metabo¬ lites.7 We set out to examine these relationships between TNF and cyclooxygenase metabolites in porcine sepsis by studying the effects of anti-TNFa MoAB pretreatment on plasma thromboxane levels. CLINICAL SIGNIFICANCE Clinical trials are presently in progress to evaluate antiTNFa MoAb as a therapeutic option in gram-negative sepsis.13 However, the effects of blocking TNFa are not fully elucidated and there is concern over potential dele¬ terious effects of blocking the normal cytokine response.8 The early TNF surge in plasma and subsequent cytokine cascade associated with cyclooxygenase and lipoxygenase
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Michigan User on 06/16/2015
metabolites of arachadonic acid represent a complex bio¬ logical response to bacterial invasion. Gathering informa¬ tion concerning the physiological and biochemical effects of TNFa blockade is central to a better understanding of the biological effects of this therapeutic option. MATERIALS AND METHODS Anti-TNFa MoAb The IgGl anti-TNFa MoAB used was purified from murine hybridoma culture harvests via cell separation, polyethylene glycol precipitation, anion exchange, and size exclusion chromatography. Purified MoAb was 99% pure with fully functional binding to human TNF. Endotoxin levels were less than 2 pg/mg of protein (limulus assay). Stabilization was performed with glycine and maltose before lyophilization. Lyophilized MoAb was kept at 4°C until the time of the experiment, when it was re¬ constituted with 20 mL of sterile water. Anti-TNFa MoAb (15 mg/kg) was administered for 10 minutes via the left internal jug¬ ular vein.
Porcine Model Yorkshire pigs weighing between 15 and 33 kg were obtained from a commercial vendor and kept in the Virginia Common¬ wealth University vivarium for 3 to 5 days before use. All animals received intramuscular penicillin G benzethine and penicillin G procaine (300 000 U each) 48 hours before use as part of preop¬ erative conditioning. Animals were preanesthetized with intra¬ muscular ketamine hydrochloride (25 mg/kg) and placed supine. Pentobarbital sodium (20 mg/kg) was administered intrave¬ nously to induce anesthesia, which was maintained with con¬ tinuous pentobarbital infusion (1 mg/kg per hour). After ade¬ quacy of anesthesia was determined, animals were paralyzed with a continuous intravenous infusion of pancuronium bro¬ mide (0.2 mg/kg per minute). Trachéal intubation was performed with a cuffed endotracheal tube (National Catheter). Ventilation was performed with a large-animal ventilator (Harvard Appara¬ tus, Boston, Mass) employing a 50% fraction of inspired oxygen and 5 cm H20 positive end-expiratory pressure. Vascular cath¬ eters were placed in the left common carotid artery for monitor¬ ing systemic arterial pressure and determining arterial blood gas levels and in the left external jugular vein for infusion of bacte¬ ria. An indwelling balloon-tipped pulmonary arterial catheter was inserted via the right external jugular vein and positioned via pressure monitoring in the pulmonary artery for measure¬ ment of pulmonary arterial pressure and cardiac output (Ed¬ wards COMI, Santa Ana, Calif). Cardiac index (CI) was calcu¬ lated using the following formula:
Cardiac Output/[(Body Weight20) x 0.112] A 5F lung-water catheter (model 96-020-5F, American Edwards Laboratories, Tustin, Calif) was positioned in the lower abdom¬ inal aorta via the left femoral artery for measurement of extravascular lung water (EVLW) with the thermal indocyanine green (Cardio-Green) double indicator technique.14 Systemic ar¬ terial blood sample and mixed venous samples were drawn from the distal port of the pulmonary arterial catheter for blood gas analysis using a pH/blood gas analyzer (model 1304, Instrumen¬ tation Laboratories, Lexington, Mass). Hemodynamic and blood gas variables were measured at 30-minute intervals for the first hour and every hour thereafter. Extravascular lung water mea¬ surements were taken at baseline and at 60-minute intervals un¬ til the end of the 5-hour study period. CI
=
TNF
fuged at 500g for 20 minutes at 4°C and the resulting plasma was
frozen at 20°C until time of assay. The L929 cells were seeded into flat-bottomed 96-well microtiter plates (Corning Glass Works, Corning, NY) at a density of 4 x 104 cells per well and grown to confluence overnight in Dulbecco's minimal essential medium (DMEM; GIBCO BRL, Grand Island, NY) containing 1% penicillin-streptomycin and 5% fetal calf serum (DMEM). The medium was then removed from confluent monolayers and 100 µ of DMEM containing dactinomycin (Merck, Sharp and Dohme, Westpoint, Pa) in a final concentration of 5 mg/L was added to each well. One hundred microliters of each of the fol¬ lowing was then added to selected duplicate wells containing L929 cells: (1) DMEM (0% cytotoxicity); (2) serial dilutions of recombinant TNF (5xl0~3 to 6xl0~4 U/mL) (Cetus Corp, Em¬ eryville, Calif); (3) plasma samples from different groups; (4) DMEM in blank wells without cells (100% cytotoxicity). The plates were then incubated for 20 hours at 37°C in 5% carbon di¬ oxide. The medium was removed after incubation, and the L929 cells were stained for 10 minutes with 0.5% crystal violet in 20% methanol, rinsed in water, and air dried. Optical density of each well was determined with a microplate reader (Bio-Tek EL 309) and calibrated to noncellular reagent blanks at a wavelength of 550 nm. The percentage of cytotoxicity of L929 cells was calcu¬ lated with the following formula: % of Cytotoxicity = (Optical Density of Wells With 0% Cytotoxicity Optical Density of Experimental Sample Well)/ Optical Density of Wells With 0% Cytotoxicity Tumor necrosis factor activity is expressed in units per milliliter, with one unit of TNF activity defined as 50% L929 cytotoxicity. —
—
Plasma Thromboxane B2
—
Study Groups Three groups of animals were studied. Groups 1 (n 12) and 2 (n 6) received a 1-hour infusion of live aeruginosa (5xl08 colony-forming units [CFU]/mL at 0.3 mL/20 kg per minute). Group 2 was pretreated with 15 mg/kg of intravenous anti-TNFa MoAb 15 minutes before the Pseudomonas infusion. Group 3 (n 8) received sterile saline only. All animals were studied for 300 minutes. The experimental protocol for this study was approved by the Institutional Animal Care and Use Committee of Virginia Com¬ monwealth University and adhered to National Institutes of Health guidelines for the use of experimental animals. =
=
=
Statistical
Analysis
Results are expressed as means ± SEMs. Differences were tested for significance using analysis of variance. Differences between means were analyzed using Tukey's Studentized Range Test. Results were considered statistically significant at P
