Br. J. Surg. 1992, Vol. 79. October. 1087-1 090

Prevention of cardiovascular

effects of endotoxaemia by monoclonal antibodies specific for core endotoxin

J. Reidy, 1. Wright, S. J. Boom, R . Barclay*, F. Di Padovat and G. Ramsay University Department of Surgery, Western Infirmary, Glasgow, *Scottish National Blood Transfusion Service, South- East Regional Centre, Edinburgh, UK and tPreclinica1 Research Division, Sandoz Ltd, Basel, Switzerland Correspondence to: Mr G. Ramsay, Department of

Surgery, Western Infirmary, Glasgow G I 1 6NT, UK

Passive immunization with antibody to the core region of endotoxin (core lipopolysaccharide ( L P S ) ) has been reported to reduce mortality in severe sepsis. A rat model of endotoxaemia that reproduces the hyperdynamic cardiovascular state seen in early sepsis was developed to test monoclonal antibodies specijic for core LPS. A thermodilution technique of measuring cardiac output was adapted for use in rats. Twenty-Jive animals were anaesthetized and mechanically ventilated with monitoring of central venous pressure and mean arterial pressure. Fluid replacement was adjusted to maintain the central venous pressure. Controls ( n = 10) and antibody-treated animals ( n = 5 ) showed no signijicant change in cardiac output. Animals given 0.1 mg kg- R2 endotoxin over 1 h (n = 5 ) showed a significant rise in cardiac output of 65per cent ( P < 0.01). This was abolished in rats given both antibody and endotoxin ( n = 5 ) . This study provides evidence that a monoclonal antibody against core L P S abolishes the hyperdynamic state induced by endotoxin infusion.

The sepsis syndrome with subsequent multiple organ failure continues to be a major cause of mortality among critically ill surgical patients'. Macrophage stimulation, releasing cytokines such as tumour necrosis factor and other secondary mediators, is the major cause of this syndrome*. Macrophage stimulation may be provoked by circulating endotoxin derived from a source of Gram-negative infection or, more commonly, from the patient's own gut f l ~ r a ~ Endotoxin, -~. which is known chemically as lipopolysaccharide (LPS ), is a major constituent of the outer cell wall of Gram-negative aerobic bacilli. The core region of the molecule (core LPS) is similar in structure and chemical composition in most pathogenic Gram-negative bacilli6. Passive immunization with polyclonal human antibody specific for core LPS may be beneficial in critically ill Monoclonal antibodies specific for core LPS were developed and the efficacy of these tested in a small animal model simulating the cardiovascular derangements seen in human sepsis. A whole-chest thermodilution method of measuring cardiac output in the rat was developed. Its accuracy was verified by comparison with measurements determined by an electromagnetic flowmeter. An endotoxaemic rat model was developed with which to test the efficacy of these antibodies.

Materials and methods Aninials Male Sprague-Dawley rats weighing between 250 and 450 g were used. They were kept under standard conditions of heat and light, and provided food and water ad libitum until the time of experiment. Monoclonal antibody H513.23 is a murine monoclonal immunoglobulin G,, specific for core LPS. It was produced by a collaboration between the University Department of Surgery, the Preclinical Research Division of Sandoz, Switzerland, and the Scottish National Blood Transfusion Service, Edinburgh. In vitro testing of H513.23 demonstrated specific binding to a range of clinically relevant endotoxins (unpublished data).

Cardiac ouipul nieusurenients Thermistors were salvaged from pulmonary artery flotation catheters and mounted in 2-Fr catheters (Portex, Hythe, UK). After induction of anaesthesia, under aseptic conditions, a ventral incision was made in the neck, and the right common carotid artery and external jugular veins isolated and controlled. A thermistor catheter was inserted into the common carotid artery, advanced into the aortic root and secured. A 3-Fr polyethylene catheter ( Portex) cut to a length of 20 cm was inserted through the external jugular vein into the right atrium. Cardiac output measurement was made by injecting a bolus of 155 pI saline at room temperature into the right atrium. Temperature change registered by the thermistor was analysed by a personal computer program written in Turbo-Pascal 4 (Borland, Scotts Valley, USA) to produce a thermodilution curve. The area under the curve allowed calculation of the cardiac output.

Meusurement of arterial pressure and central Denous pressure Mean arterial pressure was measured through a 3-Fr polyethylene catheter in the right femoral artery. Central venous pressure was measured from the right atrial line. E.\-permnient I Nine rats were anaesthetized by intramuscular injection of 0.3 ml kg- ' Hypnorm (fentanyl 0.315mg ml-' and fluanisone 10 mg m1-I; Veterinary Drug Company, York, UK). Vascular cannulas and a right femoral venous line were placed as described. Tracheostomy was performed and the animals mechanically ventilated with room air. Through a midline sternotomy the aortic arch was identified, cleaned and a 2-mm electromagnetic flow probe (Blood Flow Meter, Model SP2202; Could Statham, Hato Rey, Puerto Rico) placed around it between the innominate and left common carotid arteries. A ligature was placed around the origin of the innominate artery ensuring that all the cardiac output, less the coronary circulation, passed through the aortic arch. The chest wall was closed around the lead of the flow probe. A variety of high and low output cardiovascular states were then induced by controlled haemorrhage via the arterial line, reperfusion of shed blood through the femoral vein line and infusion of dopamine hydrochloride (Abbott, Queensborough, U K ) at a rate of 10-20 pg kg- minSimultaneous determinations of cardiac output measured by the electromagnetic flowmeter and whole-chest thermodilution method were taken. The animals were killed at the end of each procedure.

'

Presented to the Association of Surgeons of Great Britain and Ireland in Oxford, UK, April 1991

OOO7-1323/92/101087-04 0 1992 Butterworth-Heinemann Ltd

'.

1087

Cardiovascular effects of endotoxaemia: J . Reidy et al. Experiment 2 Twenty-five rats were anaesthetized and underwent tracheostomy and mechanical ventilation with a mixture of 1 per cent halothane, 69 per cent nitrous oxide and 30 per cent oxygen. Thermistor and vascular lines were placed as before. All the animals were infused with gelatin solution (Haemaccel; Hoechst, Hounslow, U K ) and fluid replacement adjusted to keep central venous pressure between 3.5 and 4.5mmHg throughout the experiment. Body temperature was maintained between 37 and 38°C using a radiant heat source. All animals were allowed to recover for a period of 1 h, at which time measurements of mean arterial and central venous pressures were taken. Cardiac output was measured by whole-chest thermodilution; at least two measurements were taken at each time point. This first observation point was designated as T = 0 min. Readings were repeated every 60 minup to 240 min. The animals were then killed.

rats. These received 10 mg kg-' H513.23 at T = - 15 min. The volume of antibody solution was approximately 0.3 ml, which was infused slowly over a period of 5 min. Group 4 ( n = 5) contained antibody- and endotoxin-treated animals. These were pretreated at T = - 15 min with 10 mg kg-' H513.23 followed by an infusion of 0.1 mg kg-' R2 endotoxin from T = 0 to T = 60 min. Statistical methods Correlation between the electromagnetic flow rates and thermodilution cardiac output measurements were analysed by calculation of the coefficient of correlation and numerical agreement with the method of Bland and Altman". For comparison of cardiac output between the experimental groups in experiment 2, two summary statistics were chosen" and compared using Student's unpaired t test. The summary statistics used were maximum change in cardiac output and the mean area under the curve of cardiac output versus time for each group. This latter figure was an estimate of total cardiac output in each group over the observation period.

Experimental groups Animals were divided into four groups. Group 1 ( n = 10) were controls and received fluid resuscitation only. Group 2 ( n = 5) contained LPS-treated animals which received an infusion of 0.1 mg kg- R2 LPS (extracted from an R2 core type rough mutant of Escherichia coli) from T = 0 to T = 60 rnin. Group 3 ( n = 5 ) consisted of antibody-treated

'

-

Results Experiment 1 A total of 258 paired estimations of cardiac output were obtained and plotted on a scattergram (Figure 1 ). There was

300

7

I

.-c

-

250

I

0

Y

200 1

X

U

.-c

150

L

c 0)

; m

a

k!

-n .-

5

100

" I $ -40

sot ~

V

0

,

1

50

100

1

150

1

200

1

250

1 50

300

100

E l e c t r o m a g n e t i c flow i n d e x ( m l kg-l m i n - ' )

Figure 1 Scattergram showing cardiac parameter ( thermodilution cardiac output) plotted against electromagnetic flow rate in the rat aortic arch. Values are indexedaccording to body-weight. r = 0.98, P < 0,001

..

E

200

250

300

Figure 2 Bland-Altman plot showing the difference between each pair of thermodilution cardiac output and electromagnetic flow rate readings plotted against their mean value. The mean of the differences (bias) is 3.27mlkg-' min-' (normal range ( 2 s . d . )20~97rnlkg-'min-')

t

t

8or+--

120

120

L

E E

I 0

I

I

I

I

60

120

180

240

0

60

Time ( m i n )

120

180

240

Time (min)

a

b 120

1

40t I

0

C

150

Mean o f v a l u e s (ml kg-' min-' )

120

I

I

I

I

60

120

180

240

Time (min)

1

I 0

d

I

I

I

I

60

120

180

240

Time (min)

Figure 3 Mean(s.e.m.) arterial pressure ( M A P ) plotted against time. a Group I control animals ( n = 10). b Group 2 lipopol~saccliuride-treated animals (n = 5 ) . c Group 3 antibodv-treated animals (n = 5 ) . d Group 4 untibodv- and endotoxin-treated animals (n = 5 )

1088

Br. J. Surg., Vol. 79, No. 10, October 1992

Cardiovascular effects of endotoxaemia: J. Reidy et al.

3 2

-

100-

-

0

80 60

.-

40-

-% -. -an ._1 : o r u C

-

1

-

I

I

I

I

60

120

180

240

Time [ m i n )

C

O

E

3

60

120

180

240

Time ( m i n )

d

Figure 4 Meun(s.e.ni.) perceniuge chunge in curdiuc ouipui plotted uguinsi time. a Group I control unimuls ( n = 10). b Group 2 lipopol~sacciio~ide-ireuied animals (n = 5 ) . c Group 3 untihody-treatedanimals ( n = 5 ) . d Group 4 untihody- und endotosin-treutrduninlu1.Y ( n = 5 )

a high degree of correlation between the two methods. The degree of arithmetical agreement between the methods was analysed in a Bland-Altman plot ( F i g u r e 2 ) . The difference in measurement between each method was plotted against the mean of the two measurements; the mean(2s.d.) of the differences was 3.27(20.97) ml kg-' min-I. Esperinimt 2 The results of measurements of mean arterial pressure and cardiac output in the four experimental groups are shown in Figures 3 and 4 . Central venous and mean arterial pressures were similar throughout the procedure for all groups. The cardiac output in the control animals (group 1 ) remained close to the baseline level throughout the observation period. The animals pretreated with H513.23 (group 3 ) showed a rise in cardiac output which reached a maximum at T = 120 min, but this was not significantly different from that in the control group. The animals treated with R2 LPS (group 2 ) showed a rise in cardiac output that was maximal at 1 h. Using both peak cardiac output and mean area under the curves, the increase in cardiac output in the LPS-treated animals was significant compared with controls (maximal increase in cardiac output: r = 3.7, P < 0.01; area under the cardiac output curves: t = 2.7, P < 0.05; 13 d.f.). In thegroup pretreated with H513.23 and then administered LPS (group 4 ) this increase in cardiac output was abolished despite maintenance of normal central venous and mean arterial pressures.

Discussion The sepsis syndrome and multiple organ failure continue to be a major clinical problem in critically ill patients. Treatment of such patients with antibody specific for core LPS may confer benefit in terms of reduced mortality. Initial trials have used human polyclonal antibody derived from two different sources. Braude's group' collected serum from volunteers who had been inoculated with heat-killed J5 E. coli rough mutant (with only core LPS on their cell surface). Gaffin et screened a normal blood donor population for individuals who had a high endogenous anticore LPS antibody activity. Such blood was processed to produce an immunoglobulin fraction rich in

Br. J. Surg.. Vol. 79, No. 10, October 1992

anticore LPS antibody. Inoculation of volunteers with heat-killed organisms is no longer acceptable and the amount of antibody available from blood donors with high endogenous anticore LPS antibody is limited. The danger of transmission of infection is a further problem with these methods. Groups have sought to develop monoclonal antibodies specific for core LPSl3.14 The cardiovascular derangements in early human sepsis include systemic vasodilatation, increased cardiac output and decreased systemic vascular resistance. An accurate, reliable and reproducible method of cardiac output measurement had to be devised. The thermodilution method has been shown to be accurate and reliable in large animals and humansI5, although doubts about its accuracy in small animals have been raised. Isoyama e t d . I 6 showed that thermistors available for clinical use are suitable for measuring temperature change in the small vessels and relatively low flow states seen in the rat. In small animals it is necessary to measure the whole-chest thermodilution curve with the injection site in the right atrium and the sampling site in the ascending aorta. Hayes et a/." showed that the mechanics of heat distribution through the chest alters the shape of the thermodilution curve and makes interpretation difficult. In pilot experiments mathematical constants that correct for these factors in the computation of cardiac output were derived. By applying these constants to the cardiac output calculations these experiments demonstrate that cardiac output measurements correlate with estimates obtained by electromagnetic flowmeter measurement over a wide range of values. The whole-chest thermodilution method shows that suitably anaesthetized rats can be maintained in a steady cardiovascular state over a period of 4 h. Administration of a relatively small dose of endotoxin infused over a period of 1 h and supportedby fluid resuscitation results in a significant increase in cardiac output. This increase can be abolished by pretreatment with H513.23 monoclonal antibody specific for core LPS. The cardiovascular changes seen in this model, i.e. high cardiac output associated with normal central venous and mean arterial pressures, are similar to those seen in patients suffering from the sepsis syndrome. Since the cardiovascular changes in this model are initiated by infusion of highly purified endotoxin

1089

Cardiovascular effects of endotoxaemia: J. Reidy et al. derived f r o m a k n o w n species of Gram-negative aerobic bacillus, it will be possible t o c o m p a r e different c o m b i n a t i o n s of a n t i b o d y a n d endotoxin t o assess t h e cross-reactivity of t h e different monoclonal antibodies. I t is hoped t h a t antibodies will b e identified t h a t h a v e in uioo cross-reactivity against endotoxins derived f r o m t h e Gram-negative bacilli t h a t most c o m m o n l y colonize t h e h u m a n g u t a n d / o r are involved in infection. S u c h antibodies m a y have a useful role i n t h e treatment of t h e sepsis syndrome.

8. 9. 10.

11.

Acknowledgements This work was supported by a research grant from Sandoz.

References 1.

2. 3. 4. 5.

6. 7.

1090

Machiedo GW, Lo Verme PJ, McGovern PJ, Blackwood JM. Patterns of mortality in a surgical intensive care unit. Sury Gynecol Obsrer 1981; 152: 757-9. Maier RY. Multisystem organ failure. Arch Surg 1986; 121: 201 -8. Border JR, Hassett J. La Duca J et ul. The gut origin of septic states in blunt multiple trauma (ISS = 40) in the ICU. Ann Sury 1987; 206: 427-48. Meakins JL, Marshall JC. The gastrointestinal tract: the ‘motor’ of multiple organ failure. Arch Surg 1986; 121: 197-200. Ledingham IMcA, McCartney AC, Ramsay G , Wright I . Endotoxins as mediators. Proy Clin B i d Rrs 1988; 264: 125-34. Brade H, Brade L, Reitschel ETh. Structure-activity relationships of bacterial lipopolysaccharides (endotoxins ). Current and future aspects. Zentralbl Bakr H j y 1988; 268: 151-79. Ziegler EJ, McCutchan JA, Fierer J ct ti/. Treatment of Gram-negative bacteremia and shock with human antiserum to a mutant Esclterirhia coli. N Enyl J M e d 1982; 301: 1225-30.

12. 13.

14.

15.

16. 17.

Lachman E, Pitsoe SB. Gaffin SL. Anti-lipopolysaccharide immunotherapy in management of septic shock of obstetric and gynaecological origin. Ltotcrt 1984; i : 981 -3. Baumgartner JD,McCutchan JA, Van Melle G rt ul. Prevention of Gram-negative shock and death in surgical patients by antibody to endotoxin core glycolipid. Lancet 1985; ii: 59-63. Schedel I, Dreilchomsen U, Wentwig B et ul. Treatment of Gram-negative septic shock with an immunoglobulin preparation: a prospective randomised clinical trial. Cvit Cure Merl 1991; 19: 1104-14. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lrrriwr 1986; i : 307-10. Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. BMJ 1990; 300: 230-5. Ziegler EJ, Fisher CJ, Sprung C L r f a / . Treatment of Gram-negative bacteremia and septic shock with HA-I A human monoclonal antibody against endotoxin. N Ertgl J Mid 1991; 324: 429-36. Greenman RL, Schein RMH, Martin MA cr ul. A controlled clinical trial of E5 murine monoclonal IgM antibody to endotoxin in the treatment of Gram-negative sepsis. J A M A 1991; 266: 1097- 102. Evonuk E, Imig CJ. Greenfield W. Eckstein JW. Cardiac output measured by thermal dilution of room temperature injectate. J Appl P l ~ ~ s1961; i ~ l 16: 271-5. Isoyama T, Sat0 T, Tanaka J, Shatney CH. Measurement of cardiac output in small animals by aortic thermodilution. J Surg Rrs 1982; 33: 170-6. Hayes BE, Will JA, Zarnstorff WC, Bisgard GE. Limitations of thermodilution cardiac output measurements in the rat. Airi J Phy.Yiol 1984; 246: H754-60.

Paper accepted 4 April 1992

Br. J. Surg., Voi. 79, No. 10, October 1992