Veterinary Research Communications, 16 (1992) 453-464 Copyright ~) Kluwer Academic Publishers by - Printed in the Netherlands

ENDOTOXIN-INDUCED MICROVASCULAR INJURY IN ISOLATED AND PERFUSED PIG LUNGS B. URBAIN, P. GUSTIN AND M. ANSAY Laboratory of Veterinary Pharmacology and Ecotoxicology, Faculty of Veterinary Medicine, University of Liege, Bd de Colonster Sart-Tilman B41, B-4000 Liege, Belgium

ABSTRACT Urbain, B., Gustin, P. and Ansay, M., 1992. Endotoxin-induced microvaseular injury in isolated and perfused pig lungs. Veterinary Research Communications, 16 (6),453-464 The lungs of 13 healthy Landrace piglets were isolated, perfused and maintained in an isogravimetric state under zone III conditions. By applying vascular occlusion methods, the total blood flow resistance (Rt) was partitioned into four components: arterial (Ra), pre- (Ra') and post-capillary (Rv'), and venous (Rv). The capillary filtration coefficient ( K ) was evaluated using a gravimetric technique. A bolus of 55/zg of Eschen'clu'a coli endotoxins (LPS) per 100 g of lung was injected into the arterial reservoir of eight lungs, followed by an infusion of LPS at a rate of 55/xg per I00 g of lung per hour for 180 min. A bolus of theophyUine (85 mg per 100 g of lung weight) was injected into the arterial reservoir after the last determination of the K value. All the parameters were evaluated again when the lungs reached a new steady state• Endoto2tn reduced a slgmficant increase m Rt from 54.7 __. 7.0 at zero t~me to 184.7 _+ 44.2 cmH O mm L (100 g) 180 minutes later, which can be ascribed to the increase m Ra and t i v . Tliese haemodynamlc modfftcatlons were related to the increases in the arterial pressure and in the pressure at the distal end of the arterial segment and to the decreases in the pressure at the proximal end of the venous segment and in the blood flow. The capillary pressure and the lung weight remained unchanged• Endotoxin infusion induced an increase in the K value from 0.208 -+ 0.011 (at t=0) to 0•391 -+ 0.034 ml rain -1 (cmH O) -1 (100 g)-I (at t= 180)./~dmmlstratmn of theophylhne sxgmficantly reduced Rt, Ra, Ra and Rv towards or under the baseline values and also induced a significant increase in the lung weight and in the K value. It was concluded that the endotoxln-mduced increase m the total blood flow resistance can 16~ ascribed to a vasospasm occurring at the level of the pre- and post-capillary small vessels and that changes in the permeability of the endothelium greatly contribute to the development of the pulmonary oedema observed in endotoxaemic pigs. •

fc

•

•

.

,

e

.

•

.

.

I

.

.

.

j

.

.

.

.

-1

2

.

.

.

.

-1

•

,

.

.

.

.

.

.

•

.

.

,

2

fe

.

Keywords: endothelial permeability, endotoxins, Escherichia coli, haemodynamics, lungs, pigs

INTRODUCTION Intravenous administration of toxin from Gram-negative bacteria has been shown to provide a good model of bronchopneumonia in pigs (Olson et aL, 1986). It also reproduces the physiopathological manifestations observed during human adult respiratory distress syndrome (ARDS) and neonatal sepsis (Seeger and Suttorp, 1987). The high sensitivity of pigs to Escherichia coli endotoxin-induced acute lung injury seems to be linked to the fact that lipopolysaccharides are preferentially sequestered into the lungs by intravascular macrophages (Warner et al., 1988). Stimulation of the latter and complement activation are followed by the release of inflammatory mediators such as serotonin, histamine, platelet-activating factor, prostaglandins, thromboxane A2, leukotrienes, toxic oxygen radicals, tumour necrotic

454 factor and lysosomal enzymes from activated inflammatory cells and pulmonary tissue (Schrauwen et aL, 1985; Brigham and Meyrick, 1986; Olson et aL, 1986; Olson, 1987; Schumacher et al., 1987; Warner et al., 1988). Moreover, lipopolysaccharides may have direct toxic effects on endothelial cells (Harlan et al., 1983; Meyrick et aL, 1989). It is well known that acute respiratory failure induced by lipopolysaccharides is accompanied by bronchoconstriction, hypoxaemia, pulmonary hypertension with increased pulmonary vascular resistance and oedema (Schrauwen et al., 1985; Olson, 1987; Schumacher et aL, 1987; Olson et aL, 1988). However, the microvascular haemodynamics and the changes in endothelial permeability during endotoxaemia in pigs have not so far been investigated. The aim of the present study was to evaluate the effects of a moderate dose of endotoxins on the capillary filtration coefficient and on the partitioning of the resistance to the blood flow across the different segments of the vascular bed in isolated and perfused pig lungs. In order to determine whether the changes in the haemodynamic parameters induced by endotoxins were due to vasoconstriction, the effect of theophylline was also tested.

MATERIALS AND METHODS General procedure

Healthy Landrace piglets, of both sexes, weighing 30.0 -- 1.3 kg (n = 13) were anaesthetized with azaperone (Stresnil; Janssen Pharmaceutica, Belgium) (8 mg/kg IM) followed by ketamine (Imalgene 1000; Rhone Merieux, France) (15 mg/kg IM) and, if necessary, by thiopental (Penthotal; Ceva, Belgium) (up to 10 mg/kg IV). The methods used in this study have been described previously (Gustin et al., 1992). Briefly, the animals were heparinized (Liquemine; Roche, Belgium) (1000 IU/kg IV) and then bled through a catheter placed in the femoral artery. The blood was mixed with an artificial perfusate (Krebs - Ringer bicarbonate buffer fortified with 45% (w/v) albumin; Sigma Chemical Co, St Louis, MO, USA). The fmal haematocrit was 20%. A tracheotomy was performed and the animal was intubated and ventilated with oxygen. The thorax was opened by median sternotomy and the heart-lung block was excised. The ventricles were removed by cutting off the heart at the atrioventricular groove and the lungs were weighed. Glass cannulae with side ports were secured in the left atrium, just above the pulmonary veins and in the main pulmonary artery. The preparation was then rapidly connected to a classical recirculating perfusion circuit consisting of the following elements: an open venous reservoir, an electronic roller pump (NY-7550-62, Masterflex, Bioblock; Illkirch, France) indicating blood flow, a heat exchanger, an open arterial reservoir and a bubble-trap. Care was taken not to introduce air into the circuit. The liquid level in the arterial reservoir was kept constant by adapting the blood flow to changes in Rt. In this constant-level setup the height of the reservoirs determines the arterial and venous pressures (Townsley et al., 1986). Arterial pressure also depends on the vasomotor tone of the vascular bed. The lungs were ventilated with humidified air at a frequency of 15 cpm, a tidal volume of 10 ml/kg of body weight and an end expiratory pressure (Pc) of 3 cm H20. The blood in the venous reservoir was aerated with 0 2 and CO 2. The pH of the arterial blood was continually checked and adjusted to 7.35-7.45 by adding 1 mol/L

455 NaHCO 3 to the perfusate or by increasing the CO 2 flow. The partial pressures of oxygen and carbon dioxide were maintained within physiological ranges: 100-140 mmHg and 25-35 mmHg, respectively. The blood temperature was 37°C. The arterial (Pa) and venous (Pv) pressures were measured using two thin closed-end fluid-filled catheters with side holes which were introduced into the main pulmonary artery and into the left atrium through the side ports of the cannulae and connected to pressure transducers (P23 ID and P50 Gould; Brussels, Belgium). The pressures were zero referenced at the lung hilus. Pa and Pv were adjusted to give a blood flow of about 20 ml min -1 kg -I body weight. Equilibration was obtained when the lungs became isogravimetric, i.e. neither gaining nor losing weight, under zone III conditions (Pa > Pv~ > Pe), which are characterized by the opening of the maximum number of vessels. The lungs were placed on a balance (L2200 Sartorius; Vanderheyden, Brussels, Belgium) in a plastic~bag (Yoshimura et al., 1989). They were periodically hyperinflated to prevent atelectasis and were covered with a plastic wrap to prevent evaporative fluid loss. All parameters were recorded on a Gould recorder (TA 2000; Brussels, Belgium).

Experimental measurements The total pressure gradient across the vascular bed (Pa - Pv) was partitioned into pressure gradients across the arterial, venous and middle segments by the arterial and venous occlusions method proposed by Hakim et al. (1982) in dogs and recently adapted to pigs (Gustin et aL, 1992). Briefly, inflow and outflow occlusions were successively performed using two electromagnetic valves placed around the arterial and venous tubes near to the cannulae, for a brief period (2-3 s) at end-expiratory pressure. When the inflow was stopped, the Pa fell rapidly by an amount equal to the pressure gradient across the arteries (APa = Pa - Pa') and then fell more slowly due to the vascular and lung tissue compliance. Clamping the outflow induced an abrupt increase in Pv equal to the pressure gradient in the veins (z~d~v= Pv' - Pv) followed by a slow rise in Pv. Pa' and Pv' are defined as the pressures at the distal end of the arterial segment and the proximal end of the venous segment, respectively. Detailed tracings of pressure changes after arterial and venous occlusions have been published previously (Gustin et al., 1992). The middle segment, which includes the capillaries, small arteries and small veins, was divided into pre- and post-capillary components by means of the double occlusion method which allows measurement of the capillary pressure (Pc) (Dawson et al., 1982). Simultaneous occlusion of both the inflow and outflow for a short period of a few seconds induces an increase and a decrease in Pv and Pa respectively until Pa and Pv equilibrate to the same pressure, which is a good estimation of Pc. The total resistance to blood flow (Rt) was calculated by dividing (Pa - Pv) by the blood flow (Q) when the lungs were in a steady state. The segmental resistances (Ra, Ra', Rv', Rv) were calculated from the following equations: Ra

= (Pa - Pa')/O_

(arterial resistance)

Ra'

= (Pa' - Pc)l(2

(pre-capillary resistance)

456

Rv'

= (Pc - Pv')/Q

(post-capillary resistance)

Rv

= (Pv' - Pv)/Q

(venous resistance)

with Rt

=

Ra+Ra'+Rv'+Rv

Estimation of capillary filtration coefficient The capillary filtration coefficient (Kf c) was measured by increasing the Pa and the Pv, usually by 10-15 cmH20 for 8-10'min. Pc was estimated before and at the end of this period in order to determine the difference in hydrostatic pressure (APc) inducing fluid filtration across the endothelium. The amount of fluid that f'dtered across was calculated from the tracing of the weight. At first the lungs gained weight rapidly for one or two minutes following the increase in vascular pressures due to congestion. Later, the weight of the lungs increased more slowly due to further fluid filtration. The slope of the slow phase of weight increase (AW/~t) was plotted on a semilogarithmie scale as a function of time and extrapolated to zero time. This value was divided by ~Dc to give Kf,c expressed in millilitres (the fdtrate was assumed to have a density of 1.0) per minute per cmH20 per 100 g lung weight. Further details of this method have been published previously (Drake and Gabel, 1981; Sumita et aL, 1989; Gustin et al., 1992).

Experimental protocol All the parameters were measured when the lungs were in a steady state (t = 0). The resistance values were calculated from two or three occlusions performed successively over a few minutes. In the endotoxin-treated lungs (n=8), a bolus of 55 /zg of endotoxins (LPS) (E. coli) 0 127: B8; Difco, Belgium) per 100 g of lung was injected into the arterial reservoir just after measurement of the baseline values. Administration of the bolus was followed by infusion of 55/zg of LPS per 100 g of lung per hour for 180 min. The endotoxins (E. coli 0 127: B8; Difco, Belgium) were dissolved in sodium chloride solution (0.9%) (25 /zg/ml) on the day of the experiments. This procedure, i.e. a bolus followed by an infusion, was adapted from that used in vivo in piglets by Olson (1987), who infused LPS at 5/zg kg- 1 h- 1 for one hour followed by infusion at 2 /zg kg-1 h -1. Preliminary studies showed that the dosage used in this study gives a progressive and significant response without inducing the immediate and extensive damage which might result from larger doses. In the control group (n =5), sodium chloride solution (0.9%) was injected into the arterial reservoir at the same rate as in the endotoxin-treated lungs. The occlusions were performed every 30 min for 180 min and the capillary filtration coefficient determination was repeated at the end of this period (t=180). In 5 of the 8 endotoxin-treated lungs, a bolus of theophylline (Theophylline Bruneau; Delalande, Paris, France) (85 mg/100 g of lung weight) was injected into the arterial reservoir after the last determination of the Kf, c value. Measurements of the resistances and of the capillary f'dtration coefficient were performed after a new steady state was obtained (20-30 min). Samples of lungs were prepared for histological examination.

457

]

~-,~.

~'~,~

~,...,

+! +I L"',-~

+I +I ~

+I +I O0

.

~.~, +1 +1

+1 '4-1

+1 '+1

+1 +1

~.~

e4

o~ +I

+I +I

+I +I

+~ +~

+I +I

+I +I

+I +I

@

~,-,=

%

m

II ~;.

~ +I

+I +I

+I +I

+1

+1 +1

+i +1

~ ~ +I +I

~ +I +I

+I +I

+I +I

4"1 +1

÷1 41

+1 +1

+1 +I

'+I 41

41 41

'+I +I

+I +I

+t +l

+I +t

+1 +l

+I +I

+I -I-I

+I +I

II

~,~

5 8~ c~

+I ~ +1

~ +1 +1 ~ L",-

+1 L".,- 4-I

z °

0

+I

+{ +I

+I +I

+{ +I

~

0 0

II

0

~

~

~ ~.e=='~

o~

o.

o~

"7

e=¢..~ , ~

,.,=

.~

458 T A B L E II Mean values __. SEM a of the total (Rt), arterial (Ra), pre- (Ra') and post- (Rv') capillary, and venous (Rv) blood flow resistances, of the arterial (Pa), capillary (Pc), and venous (Pv) pressures, of the blood flow (Q), of the lung weight (LW) and of the capillary ftltration coefficient (Kfc) in endotoxin-treated lungs (n = 5) at zero time (t = 0), after 3 h of perfusion (t = lg0) and after administration of theophylline (Theo) t'

Parameters a

t=0

t = 180

Theo

Rt

60.4 - 9.1

212.3 _ 42.3d

38.5 _ 15.6c

Ra

12.2 _ 1.1

19.6 +_ 3.2

9.3 + 1.6c

Ra'

18.4 +- 4.8

118.3 _+ 27.9d

19.5 _+ 12.4c

Rv'

22.6 -+ 4.9

66.5 + 13.1d

4.5 + 1.7c'~

Rv

6.9 +- 1.1

7.5 + 2.0

5.1 +-. 0.9

Pa

18.8 _+ 2.2

29.6 _+ 2.1~

13.4 +_ 2.5c'd

Pc

12.8 -+ 1.6

12.7 _+ 1.1

10.1 _+ 2.1 c

Pv

6.4 -+ 1.3

3.5 _+ 1.7

7.9 -+ 2.2

622 _+ 66

433 _+ 63d

622 _+ 28 c

LW

480 ___37

498 +_ 60

655 +_ 57c'd

Kf,c

0.199 _+ 0.010

0.365 __. 0.40d

0.487 _ 0.037c'a

a

Units: Rt, Ra, R a ,, Rv,, Rv - cmH20 rain L - 1 (100 g)- 1 ; Pa, Pc, Pv - cmH20; O - ml rain-l; LW - g; K. - ml rain -1 (cmH^O) -1 (100 g)-I A bolus of endotoxins (55/zg/100 g of lungs) was injected in the arterial reservoir (t= 0) followed by an infusion at a rate of 55/zg/100 g per hour for a period of 180 min (t= 180). Theophylline (85 rag/100 g lung) was administered at the end of this period (see text) CSignificantly different from values measured after 3 h of endotoxin infusion (p < 0.05) dSignificantly different from baselines values (t = 0) Co< 0.05)

Statistical analysis Total data were analysed by two-way analysis of variance to determine whether significant differences exist between the mean values obtained within and between the

459

two groups. When the ANOVA was significant, a Student's t-test for paired or unpaired data was used for the comparison of the means. P values less than 0.05 were considered to be significant. All results are expressed as the mean _+ SEM.

RESULTS The evolution of the haemodynamic parameters, the capillary fdtration coefficient and the lung weight are presented in Table I and Figures 1 and 2. The baseline values of Rt, Ra, Ra', Rv' and Rv were 55.3 + 6.1, 7.4 -+ 1.5, 17.4 _+ 3.1, 19.7 +- 3.2, 10.8 -+ 2.4 cmH20 min L -1 (100 g)-I in the control group and 54.7 _+ 7.0, 12.0 _+ 1.1, 16.3 + 3.1, 20.4 _+ 3.9, 5.9 -+ 0.8 in the endotoxin group. No significant differences existed between the groups, except for the arterial resistance (Ra) which was slightly but significantly higher in the endotoxin group than in the control group. This difference, observed throughout the experiments, was independent of endotoxin administration. The Kf c values were 0.207 -+ 0.020 and 0.208 _+ 0.011 ml min -1 (cmH20) -1 (100 g)-I in the ~ontrol and in the endotoxin group, respectively. In the former, the values of all parameters remained virtually unchanged during the 3 h of perfusion following the equilibrium period. No significant effect due to time was observed.

250 -

A

200 A A 150

,

*

Rt 100

50

0

I

0

J

30

I

60

i

90

120

150

180

time (min.)

Figure 1. The evolution of the mean values -+ SEM of the total blood flow resistance (Rt) in control lungs (solid columns) (n =5) and in endotoxin-treated lungs (hatched columns) (n = 8)./x and * indicate a significant difference (p < 0.05) with respect to the corresponding values measured in the control group and to the baseline values measured at zero time, respectively. Units: cmH20 rain L -1 (100g) -1

460 A

25

I40"i

~

120 20

*

T

A

15"

80

Ra'

Ra

60

,

10"

40 5"

20 0 0

30

60

90 time

120 (rain)

0

30

60

30

60

"

80 "

90 120 time (mtn)

150

180

90 120 time (rain)

150

180

20 18 t6 14

70" 60 ~0'

12

,

Rv~ 3 0 " 4 0 ~ 2 0 . 1 0 .

Rv

0

11

0

0

30

60

90 120 time (rain)

150

180

,

0

,

Figure 2. The evolution of the mean values +- SEM of the arterial (Ra), pre- (Ra') and post- (Rv') capillary and venous (Rv) blood flow resistances in control lungs (solid columns) (n=5) and in endotoxin-treated lungs (hatched columns) (n=8). A and * indicate a significant difference (p