Altered pulmonary microvascular reactivity after total cardiopulmonary bypass

Altered pulmonary microvascular reactivity after total cardiopulmonary bypass Pulmonary vascular resistance is frequently elevated after cardiac operations in which cardiopulmonary bypass is used. In our study of the possible contribution of altered pulmonary microvascular reactivity to this condition, sheep were heparinized, cannulated via the aorta and right atrium, and placed on total cardiopulmonary bypass. Mter 90 minutes of total cardiopulmonary bypass and pulmonary arterial occlusion, the sheep were removed from cardiopulmonary bypass, and their lungs were perfused normally for 60 minutes. Noninstrumented animals were used as controls. To evaluate the effect of 90 minutes of extracorporeal circulation without reduced pulmonary perfusion, we studied additional sheep after they underwent right heart bypass with a pump-oxygenator, Pulmonary microarterial vessels (130 to 230 #Lm in diameter) from each group were examined in vitro in a pressurized (20 mm Hg), no-flow state with video microscopic imaging and electronic dimension analysis. Mter preconstriction of vessels with the thromboxane Az analog U46619 by 30 % to 40% of the baseline diameter, vasoactive drugs were applied extraluminally. Serotonin caused control microvessels to dilate. In the presence of the nitric oxide synthetase inhibitor NG-methyl-L-arginine, this was converted to a significant contractile response. Acetylcholine alone had minimal effect on control vessels. However, in the presence of the cyclooxygenase inhibitor indomethacin, acetylcholine caused a significant relaxation response. After total cardiopulmonary bypass and pulmonary reperfusion, pulmonary microvessels contracted significantly when exposed to acetylcholine and serotonin, compared with respective control responses. Both these contractile responses were inhibited in the presence of indomethacin. Endothelium-independent responses to sodium nitroprusside and U46619 and dilation responses to adenosine were not altered after cardiopulmonary bypass. Extracorporeal circulation with continued pulmonary arterial perfusion (right heart bypass group) had no effect on microvascular responses. In conclusion, total cardiopulmonary bypass with associated reduced pulmonary perfusion causes significant alterations of endotheliumdependent pulmonary microvascular responses because of the increased release of a constrictor prostanoid substance and possibly because of reduced release of endothelium-derived relaxing factor. (J 'fHORAC CARDIOVASC SURG 1993;106:479-86)

Tajammul Shafique, MD, Robert G. Johnson, MD, Hai Bin Dai, MD, Ronald M. Weintraub, MD, and Frank W. Sellke, MD, Boston, Mass.

From the Division of Cardiothoracic Surgery, Department of Surgery, the Charles A. Dana Research Laboratory, Beth Israel Hospital, and Harvard Medical School, Boston, Mass. Supported by National Heart, Lung, and Blood Institute grant HL46716 and American Heart Association, Massachusetts Affiliate grant 13-501-912. Received for publication May 19, 1992. Accepted for publication Aug. 31, 1992. Address for reprints: Frank W. Sellke, MD, Division of Cardiothoracic Surgery, Dana 905, Beth Israel Hospital, 330 Brookline Ave., Boston, MA 02215. Copyright © 1993 by Mosby-Year Book, Inc. 0022-5223/93 $1.00 +.10

12/1/42167

Ateration in vascular resistance and mismatched ventilation and perfusion in the lungs are recognized components ofpulmonary dysfunction after cardiac operations requiring cardiopulmonary bypass (CPB). Although this is controversial, the primary site of vascular resistance in human pulmonary circulation is thought to be at the level of microarterial vessels less than 300 #Lm in diameter. 1 Vascular reactivity and tone are determined by the interaction between neurohumoral substances, vascular smooth muscle, and endothelium. The role of endothelium is particularly significant in influencing pulmonary vascular smooth muscle contractile responses in small intrapulmonary vessels.i

479

4 8 0 Shafique et al.

Total CPB, as it is used clinically, results in complete diversion of pulmonary bloodflow in the absenceof ventilation. Although the oxygenrequirement of pulmonary tissueis not great, a state of reducedpulmonaryperfusion followed by a restorationof normal perfusion mightcause significantinjury to the lung. Furthermore, extracorporeal circulation itself is known to be associated with systemic inflammatory reactions, such as complementactivation;' whichmay havea detrimental effecton the lungs. The endothelium of systemic microvessels has been shown to be sensitive to ischemia-reperfusion injury,4,5 resulting in alteration in vascular responses to the endothelium-dependent neurohumoral agents. The in vitro vasomotoractivityof pulmonary microvasculature under normal conditions and after CPB has not been characterizedand may havesignificantimpact on cardiopulmonary function. The purposeof this study was to examinethe effectsof total CPB with its associatedlack of pulmonary arterial perfusion, on pulmonary microvascular responses and to examine the possible biochemical pathways involved in the alteration of these responses. Methods Experimental preparation. Female Dorset-Rambouillet crossbred sheep weighing 25 to 35 kg (mean 29 ± 0.6 kg) were anesthetized with intravenous a-chloralose (80 mg/kg of body weight) and urethane (500 mg/kg). The animals were grouped into three categories. Total CPR group. In the total CPB group (n = 12), animals were tracheally intubated and their lungs mechanically ventilated (Hatvard Apparatus, Inc., Millis, Mass.). Arterial blood gas and pH were measured at regular intervals during the experiment (pH blood gas analyzer 1306; Instrumentation Laboratory, Inc., Lexington, Mass.) and maintained within physiologic limits (pH 7.35 to 7.42, oxygen tension> 100 mm Hg to <300 mm Hg, carbon dioxide tension <45 mm Hg to>35 mm Hg) by adjusting the inspired oxygen fraction (Fi02), ventilation rate, and tidal volume. Systemic blood pressure was monitored by percutaneous cannulation of the femoral artery. A left thoracotomy was performed, and, after systemic heparinization (400 Il.I/kg), cannulas were placed in the right atrium and the aorta. The extracorporeal circuit consisted of a standard roller pump (Cardiovascular Instrument Corp., Wakefield, Mass.) and a bubble oxygenator (Bentley Bio-2; Baxter Healthcare Corp., Irvine, Calif.). An arterial filter (Bentley AF-1025; Baxter Healthcare) was inserted into the circuit distal to the roller pump. The circuit was primed with Ringer's lactate solution (25 nil/kg). Animals were supported by total CPB. After initial stabilization, the pulmonary artery was clamped to ensure total absence of blood flow through the pulmonary artery, and ventilation was discontinued. A ventricular vent placed via the left atrial appendage ensured pulmonary venous decompression. Pump flow was adjusted from 80 to 100 ml/kg per minute to maintain aortic pressure between 50 and 75 mm Hg. After 90 minutes of total CPB, ventilation was reestablished, and the pulmonary artery clamp was removed.

The Journal of Thoracic and Cardiovascular Surgery September 1993

The lungs were reperfused for 60 minutes, and normal pulmonary circulation was reestablished by increasing cardiac filling. Sheep were removed from CPB bypass in all experiments, Inotropic and vasoactive agents were not used in any animals to support the circulation. Mean aortic and pulmonary arterial pressures were recorded before and after CPB. Ratios of these pressures were used as an index of the resistances of the pulmonary and systemic vascular beds. Segments of the lung were excised from multiple peripheral sites, and microvessels were obtained for in vitro studies. Right heart bypass group. In the right heart bypass group (n = 2), after anesthesia, we followed the same procedure used in the total CPB group, but, instead of the aorta, the pulmonary artery was cannulated. Thus the extracorporeal circuit provided nonpulsatile flow through the lungs. As in the total CPB group, oxygenation was accomplished with a bubble oxygenator (Bentley Bio-2; Baxter Healthcare), and the lungs were not ventilated during the procedure. Right heart bypass was continued for a period of 90 minutes at a flow rate of 80 to 100 ml/kg per minute. Arterial blood gas and pH were measured at regular intervals during the experiment and maintained within physiologic limits previously given. At the conclusion of this period, peripheral segments of the lung were excised, and microvessels were obtained for in vitro studies. Control group. In the control group (n = 15), after anesthesia, animals were heparinized, a sternotomy was performed, peripheral segments of the lung were excised, and microvessels were obtained for in vitro studies. All animals were killed by exsanguination. The segments of the lung were immediately placed in cold (1 0 to 4 0 C) Krebs' buffer solution of the following composition: NaCl, 118.3 mmoljL; KCl, 4.7 mmol/L; CaCh, 2.5 mmol/L; MgS04, 1.2 mmol/L; KH2P04, 1.2 mmoljL; NaHC03, 25 mmol/L; glucose, 11.1 mmol/L, Animals were cared for in accordance with the guidelines established by the Beth Israel Hospital's Committee on Animal Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication No. 86-23, revised 1985). In vitro pulmonary microvascular studies. Microarterial vessels were carefully dissected from lobular branches of the pulmonary artery with a lOX to 60X dissecting microscope (Olympus Optical, Tokyo, Japan). These vascular segments varied in size from 130 to 230 ~m in diameter and 1 to 2 mm in length. Vessels were placed in an isolated Plexiglas acrylic plastic organ chamber, cannulated with dual glass micropipettes measuring 30 to 80 ~m in diameter, and secured with 10-0 nylon monofilament suture (Ethicon, Inc., Somerville, N.J.). Oxygenated (95% O 2, 5% CO 2) Krebs' buffer solution warmed to 370 C was continuously circulated through the organ chamber and a reservoir (total volume 100 ml). The vesselswere pressurized to 20 mm Hg in a closed, no-flow state with two burette manometers filled with Krebs' buffer solution and maintained at the same level. With an inverted microscope (40X to 2oox, Olympus IMT-2; Olympus Optical) connected to a video camera, the vesselimage was projected onto a black and white television monitor (Panasonic; Matsushita Electric Industrial Co. Ltd., Osaka, Japan). A video electronic dimension analyzer (Living Systems Instrumentation, Burlington, Vt.) was used to measure internal lumen diameter. A pressure transducer measured distending pressure via a sidearm cannula immediately

The Journal of Thoracic and Cardiovascular Surgery Volume 106. Number 3

proximal to one of the micropipettes. Measurements were recordedwith a strip chart recorder (Western Graphtec, Irvine, Calif.). Study protocol. After equilibration of the vesselfor at least 30 minutes,the vesselwas precontracted with the thromboxane A zanalogU466l9 by 30%to 40%of the baselinediameter. This degreeof precontraction was chosen because it is approximately 70% of the maximum contraction induced by either KCl or U466l9 and allowsvessels to respond by either further contraction or dilation. U46619 was chosen as the precontraction agent becauseit producesconsistent, sustained contraction in the pulmonary circulation. After stabilization of the precontracted diameter, acetylcholinechloride, serotonin, adenosine, or sodium nitroprussidewas applied extraluminally. In addition, contraction responsesto U466l9 were examined without the addition of other agents. The order of administration was random. Two to four vessels were examined from each animal, and one to four interventions were performed on each vessel. Studies with indomethacin and NG-methyl-L-arginine. In selectedexperiments, the cyclooxygenase inhibitor indomethacin (10 /Lmol/L) or the nitric oxide synthetase inhibitor NO. methyl-L-arginine (L-NMMA; 30 /Lmol/L) was added to the organ chamber reservoir at least 15 minutes before the administration of serotonin or acetylcholine. Both control vessels and vessels subjectedto total CPB were examined. Only one of these blockingagents was used for each vessel. Drugs. Acetylcholine chloride, serotonin, adenosine, the thromboxaneA z analog U46619, sodium nitroprusside, indomethacin, and L-NMMA were obtained from Sigma Chemical Company, St. Louis, Mo. Acetylcholine chloride, serotonin, sodiumnitroprusside,L-NMMA, and adenosine were dissolved in ultrapure distilledwater. U46619 was dissolved in ethanol to make a 10 mmol/L stock solution. Indomethacin was dissolved in minimal ethanol to make a 20 mmol/L stock solution. All stock solutionswere stored at -20 0 C. Alldilutions were prepared daily. Data analysis. Microvascular responses to acetylcholine chloride, serotonin, adenosine, and sodium nitroprusside are expressed as the percentage of relaxation of the U46619induced constriction. Contraction responses to U466l9 alone are expressed as the percentage of contraction of the baseline diameter. Valuesare expressedas mean ± standard error of the mean. Responsesto each drug at each concentration were compared by means of analysis of variance. Whenever significance was indicated, Fisher's least significantdifference test was used to determine significance between groups. Statistical significance was assumed when p < 0.05. Results Before CPB, the ratio of mean pulmonary artery pressure to mean aortic pressure was 0.09 ± 0.01 (mean pulmonary artery pressure 10 ± I, mean aortic pressure 107 ± 4). After 90 minutes of CPB, this ratio increased to 0.37 ± 0.06 (p < 0.05) on pulmonary reperfusion, suggesting an increased pulmonary vascular resistance after total CPB (mean pulmonary artery pressure, 32 ± 4, mean aortic pressure 86 ± 5). MAoP during CPB was 59 ± 4 mm Hg. Vessel characteristics. Pulmonary microvessels mea-

Shafique et al. 48 1

30

z

0

• •

20

---

~

-e ><

C -'

~

10

--+--

W

II:

~

zw

0

CONTROL INDOMETHACIN L-NMMA

U II:

w

Do

-10 -20 -1---.-~--,-~-.-~.,....~-+-9.0 -8.0 -7.0 -6.0 -5.0 -4.0 LOG MOLAR ACETYLCHOUNE

Fig. 1. Responsesto acetylcholinein vitro of nontreated sheep pulmonary microvessels from control group (n = 9) and microvessels in presence of cyclooxygenase inhibitor indomethacin (n = 6) or nitric oxide synthesis inhibitor L-NMMA (n = 6). Microvesselswere pressurized (20mmHg) in a no-flow state and preconstricted with U46619 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U466l9-induced contraction. Minus sign (- ) denotes contraction. *p < 0.05 compared with control.

sured between 130 and 230 JLm in internal diameter, averaging 178 ± 3 JLm in the control group and 161 ± 4 JLm in the CPB group. Percentage of preconstriction after application of the thromboxane A2 analog U46619 was 33% ± 0.4% (mean log molar U46619 -6.6 ± 0.2) in control vessels and 32% ± 0.4% (mean log molar U46619 -6.4 ± 0.3) in vessels after CPB. The degree of preconstriction was chosen so that vessels could respond with either further constriction or relaxation. Responses of control microvessels to acetylcholine. Control microvessels relaxed minimally in response to acetylcholine alone (maximum 3% ± 7%). However, in the presence of the cyclooxygenase inhibitor indomethacin, acetylcholine caused a significant relaxation of the control microvessels (maximum 23.5% ± 5.2%, p < 0.05). In the presence of the nitric oxide synthetase inhibitor L- NMMA, the control vessels constricted slightly (maximum -13% ± 6.5%, p < 0.1) in response to acetylcholine (Fig. l). Thus, acetylcholine produces the release of a constrictor prostaglandin substance and likely the release of EDRF in the pulmonary microcirculation. Responses of control microvessels to serotonin. Serotonin caused a slight relaxation response (maximum 18% ± 5%) in control microvessels. Pretreatment of the vessels with indomethacin resulted in a slight accentuation of this relaxation response (maximum 24% ± 3%, P = not significant). In the presence of the nitric oxide

The Journal of Thoracic and Cardiovascular Surgery September 1993

4 8 2 Shafique et al.

30

~ i=

S ... Ul II:

30 20

20

z

---

10

iiio o ffi Do -10

I-

--b-

CONTROL

-....

INDOMETHACIN L·NMMA

•

-8.0

-7.0

-8.0

-5.0

10 §

z

C >C C

0

... Ul II:

- t i - CONTROL

-+-

I-

Z

_

Ul

U

CPB CPB-INDOMETHACIN

II: Ul Do

-8.0

-7.0

-6.0

-5.0

0

C

Ul II:

I-

--

- t i - CONTROL

-+-

-10

Z

Ul

U -20 II: Ul Do -3lI

-7.0

-6.0

-5.0

CPB CPB-INDOIETHACIN

-4.0

LOGMOLARSEROTONIN

Fig. 2. Responses to serotonin in vitro of nontreated sheep pulmonary microvessels from control group (n = 9) and microvessels in presenceof the cyclooxygenase inhibitor indomethacin (n = 6) or nitric oxide synthesis inhibitor L-NMMA (n = 6). Microvessels werepressurized(20 mm Hg) in a no-flow state and preconstricted with U46619 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U46619-induced contraction. Minus sign (-) denotes contraction. *p < 0.05 compared with control.

0

...

-8.0

-4.0

LOG MOLAR SEROTONIN

i=

10

~

-20+-~.......~--.-~.,.---r-~---r~

-9.0

0

>C

-4.0

LOG MOLAR ACETYLCHOLINE

Fig. 3. Responsesto acetylcholinein vitro of sheeppulmonary microvessels from control group (n = 9), from total CPB group (n = 8), and from CPB group with cyclooxygenase inhibitor indomethacin (n = 4). Microvessels were pressurized (20 mm Hg) in a no-flow state and preconstrictedwith U466I9 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U46619induced contraction. Minus sign (-) denotes contraction. *p < 0.05 compared with control. §p < 0.05 compared with CPB.

synthetase inhibitor L-NMMA, serotonin caused a constriction response in the control microvessels (maximum -9.5% ± 5%, P < 0.05; Fig. 2).

Effects of total CPB on vascular responses to acetylcholine and serotonin. The contraction of pulmonary

Fig. 4. Responses to serotonin in vitro of sheep pulmonary microvessels from controlgroup (n = 9), from total CPB group (n = 8), and from CPB group with cyclooxygenase inhibitor indomethacin (n = 4). Microvessels were pressurized (20 mm Hg) in a no-flow state and preconstrictedwith U46619 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U46619induced contraction. Minus sign (-) denotes contraction. *p < 0.05 compared with control. §p < 0.05 compared with CPB.

microvessels in response to acetylcholine increased significantly after total CPB compared with the control response. After CPB, microvessels showed a mean maximum contraction in response to acetylcholine of - 26% ± 3% (p < 0.05) of the precontracted diameter as opposed to -4% ± 4.5% in the control microvessels (Fig. 3). Interestingly, this increased contractile response did not occur in the presence of indomethacin, indicating inhibition of the stimulated release of a constrictor prostaglandin substance in response to acetylcholine after CPB and reperfusion. After CPB and reperfusion, the pulmonary microvascular response to serotonin was significantly altered, compared with the control responses. Instead of causing a relaxation of microvessels (maximum 24% ± 3%), serotonin caused a significant contraction response (maximum -31 % ± 7%, p < 0.05). As in the case with acetylcholine, indomethacin partially inhibited the enhanced contractile response to serotonin after CPB and reperfusion (Fig. 4).

Responses to sodium nitroprusside, adenosine, and thromboxane A2. Sodium nitroprusside (Fig. 5) and adenosine (Fig. 6) caused equipotent relaxation of precontracted control microvessels and microvessels subjected to total CPB. Similarly, the thromboxane A2 analog U466 19, without addition of other agents, induced equipotent contraction of the control and the CPB pulmonary microvessels (Fig. 7).

The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 3

Shafique et al. 4 8 3

100

80

z

0

t=

c)(

z

0

60

c

)(

C

C

..I

\II II:

..I

z

I-

U

U

Z

\II

a::

60

\II II:

40

l-

\II

80

t=

......... ~

20

Q,

40

\II

CONTROL

CPB

II: \II Q,

CONTROL CPB

~

20

.........

o.,-::~:;::-~.,....-~.,....-~.,....-~~~....---.

o-f'-~.,....-~.,....-~.,....-~.,....-~.,....-~.,.....-.

-10.0 -9.0

·10.0 ·9.0

-8.0

-7.0

-6.0

-5.0

-4.0

LOG MOLAR SODIUM NITROPRUSSIDE

Fig. 5. Responses to sodium nitroprusside in vitro of sheep

pulmonary microvessels from control group (n = 9) and CPB (n = 6).Microvessels were pressurized (20mmHg)ina no-flow stateandpreconstricted with U46619 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U46619-induced contraction.

Effects of extracorporeal circulation with pulmonary arterial perfusion. Responses of pulmonary microvessels after right heart bypass to acetylcholine (maximum relaxation right heart bypass 10% ± 4% versus control 3% ± 7%; p = not significant), serotonin (right heart bypass 18% ± 5% versus control 16% ± 4%, p = not significant), and sodium nitroprusside (right heart bypass 69% ± 7% versus control 65% ± 6%, p = notsignificant) weresimilarto the respective responses in control vessels after 90 minutes of extracorporeal circulation without reduced pulmonary perfusion. Thus, under the experimental conditions used in the present study, it is likely that the alteration of microvascular responses to acetylcholine or serotoninare primarily due to relativepulmonary vascular ischemia and reperfusion rather than to extracorporeal circulation per se. Discussion

The majorfindingof the presentstudyisthat total CPB withassociated reducedpulmonaryperfusionfollowed by restoration of normal perfusion markedly alters pulmonary microvascular responses to the endothelium-dependent agents acetylcholine and serotonin (Figs. 3 and 4). In contrast,responses to the direct smooth-muscle vasodilator sodiumnitroprussideand the direct smooth-muscle constrictor thromboxane A2 were not affected by total CPB (Figs. 5 and 7). The microvascular response to adenosine was also unaffected by CPB (Fig. 6). Extracorporeal circulation alone with continued pulmonary perfusion (right heart bypass) did not alter microvascular responses. Although leukocyteand complement acti-

-8.0

-7.0

-6.0

-5.0

-4.0

LOG MOLAR ADENOSINE

Fig. 6. Responses to adenosine in vitro of sheep pulmonary

microvessels from control group (n = 6) and CPB group (n = 4).Microvessels were pressurized (20mmHg)ina no-flow stateand preconstricted with U46619 by 30% to 40% of baseline diameter. Agents were applied extraluminally. Responses are percentage of relaxation of U46619-induced contraction.

0

z

0

t=

~

.........

-10

CONTROL CPB

C

)(

C

..I

\II

a::

·20

I-

z\II ua:: \II Q,

-30

-40

-9.0

-8.0

-7.0

-6.0

LOG MOLAR U46619

Fig. 7. Responses to thromboxane A2 analog U46619 invitro of sheep pulmonary microvessels from control group (n = 8) andCPB group (n = 8).Microvessels were pressurized (20mm Hg) in a no-flow state without precontraction. U46619 was applied extraluminally. Responses arepercentageofcontraction of baseline diameter. Minus sign (-) denotes contraction.

vation may be produced during CPB, we found no evidence of vascular dysfunctionin pulmonary microvesselsin the right heart bypassgroup. This findingsupports the concept that the alterations in vascular responses to acetylcholine and serotonin are more likely due to pulmonary arterial flow deprivationand not to the effectsof extracorporeal circulation per se. The experimentsperformed in the presenceof indomethacin and L-NMMA provideinsight into the biochemical pathways that are altered after total CPB. Because

484 Shafique et al.

L-NMMA, an inhibitor of nitric oxide synthesis, significantly reduced the relaxation response of control pulmonary microvessels to serotonin and, to a lesser extent, to acetylcholine (Figs. 1 and 2), it is logical to assume that a significant component of the induced relaxation in response to these agents is likely due to the stimulated release of the endothelium-derived relaxing factor (EDRF).6 EDRF causes relaxation response in pulmonary vasculature by increasing cyclic guanosine monophosphate in vascular smooth muscle." McMahon, Hood, and Kadowitz'' recently reported that the pulmonary vasodilator response to vagal stimulation and release of acetylcholine is inhibited by the EDRF synthetase inhibitor N ",-nitro-L-arginine methyl ester in the cat. The possible role of prostaglandin substances was also examined in the present study. The response of control pulmonary microvessels to acetylcholine was converted from a minimal net response to a significant relaxation response in the presence of indomethacin (Fig. I). In addition, the relaxation response to serotonin in control vessels was slightly increased at higher concentrations of serotonin when the studies were performed in the presence of indomethacin (Fig. 2). In addition to causing the release of EDRF, acetylcholine and serotonin appear to cause the release of a constrictor prostaglandin substance, which may be inhibited by indomethacin, in control pulmonary microvessels. Previous studies of the pulmonary circulation have demonstrated vasoconstriction? or dilatation 10in response to various prostaglandin substances. Furthermore, cyclooxygenase-dependent pulmonary vasopressor responses to acetylcholine have been observed in the rabbit.P- II Recently, Shirai, Ninomiya, and Sada12 demonstrated angiographically that the in vivo constriction of the small pulmonary arteries of the rabbit (internal diameter 100 to 1000 }tm) in response to vagal nerve stimulation or exogenous acetylcholine is mediated by thromboxane A2-prostaglandin endoperoxide receptors and by muscarinic receptors. Furthermore, the pulmonary vasoconstriction mediated by the muscarinic receptor is mainly due to thromboxane A 2-prostaglandin endoperoxide generation. Increased concentration of thromboxane after CPB has been demonstrated.P but the altered effect of neurohumoral substances on the pulmonary microcirculation after CPB has not been determined. After total CPB and reperfusion, both acetylcholine and serotonin caused contraction of pulmonary microvessels (Figs. 3 and 4). In the presence of indomethacin, these contraction responses were markedly reduced, indicating that the stimulated release of a constrictor prostaglandin substance is facilitated after total CPR Because indomethacin did not completely normalize the response to

The Journal of Thoracic and Cardiovascular Surgery September 1993

serotonin, it is likely that a portion of these increased contractile responses is due to the reduced stimulated release of EDRF. This is consistent with our observation in control microvessels in which the EDRF mediation of responses to serotonin seemed quantitatively greater than that to acetylcholine (Figs. 1 and 2). The endotheliumindependent direct smooth-muscle relaxant sodium nitroprusside and the constricting agent thromboxane A2 analog U46619 did not show a differential effect on pulmonary microvessels after ischemia and reperfusion (Figs. 5 and 7). Thus, it is unlikely that the findings concerning the altered effects of acetylcholine and serotonin in the present study are due to any change in smooth muscle relaxation or contractile function. It is more likely that the observed contraction responses after CPB are due to alteration in the endothelium and reflect an increase in the stimulated release of a constricting prostaglandin and reduced stimulated release of EDRF. Biaggioni and associates 14 recently reported that adenosine produces pulmonary vasoconstriction in sheep through thromboxane A2 release in vivo. The mechanism of adenosine in different vessel types and the question of whether it is endothelium dependent or endothelium independent are controversial. In the present study, adenosine produced only vasodilation of isolated sheep pulmonary arterial microvessels (Fig. 6) in both control and extracorporeal circulation groups. With the exception of possible strain differences, differences in oxygen tension, or extravascular influences as seen in the in vivo examination in Biaggioni's study, we do not have an explanation for the conflicting results. Although the mechanism of alteration in endothelial function was not directly examined in the present study, it probably involves oxygen free radicals. Although bronchial flow was maintained with this preparation, the adequacy of oxygen supply in the absence of pulmonary arterial flow has not been established. Reduced endothelium-dependent relaxation has been observed after moderate hypoxia in the large pulmonary vessel rings of the rabbit'> and after chronic hypoxia in the pulmonary circulation ofthe rat, 16 Similar observations have been made after ischemia and reperfusion in large arterial and microvascular preparations in coronary arteries.v 5,17 Oxygen free radicals are known to be generated from pulmonary endothelial cells during reoxygenation after hypoxia'f and have been found to reduce endothelium-dependent relaxation, at least partially, through the enhanced degradation of EDRF. 19.20 In addition to the role of oxygen-derived free radicals, the release of cytotoxic enzymes and substances that increase leakage of the endothelial layer may contribute to vascular dysfunction after ischemia and reperfusion.

The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 3

Ratliffand associates-' observed that ultrastructural lung injury induced by CPB was related to pulmonary infiltration of polymorphonuclear neutrophils and the duration of CPB. Chenworth and associates' concluded that transpulmonary leukosequestration after CPB was related to the aortic crossclamping time and considered complement activation as a possible cause of this phenomenon. Horpacsy and associates-i demonstrated hypoxic damage sustained by unventilated lungs during clinical CPB by measuringlysosomal enzymereleaseand related postoperative ventilationrequirements to the level of these enzymes. Thus, the release or activation of lysosomal enzymesduring CPB and other factors,suchas complement activation and leukocyte"plugging" of pulmonary arterioles, may be a major factor in postbypass pulmonary parenchymal and vascular dysfunction. It is possible that the increase in the pulmonary-to-systemic pressure ratio and vascular dysfunction may have been partially due to plugging of pulmonary arterioles and activation of leukocytes. However, in the right heart bypass group, vascular dysfunctionwas not encountered inthe studiesoflimited vessels from animalsin this group. Althoughit is not addressed in the present study, the relatively limited pulmonary perfusion during CPB may accentuate vascular dysfunction, and partial CPB may reducethe detrimental actionsof these factors on pulmonary vascular reactivity. Advantages and limitations. The present study has several advantages over previous studies that examined the effectof ischemia or hypoxiaon pulmonary vascular reactivity. First, the pulmonary microcirculation was examined in vitro. Although large pulmonary arteries mayprovide someinfluence on pulmonaryvascular resistance,it is likelythat the majority of resistanceto pulmonaryflow resides in vessels lessthan 300 ~m in diameter.' Thus, the examination of vessels in the size class used in our present study may have more clinical relevancethan previous in vitro studies of larger vessels. Also, the influence of extravascular tissue was absent in the present studybecauseexperimentswere performed in vitro. Vasculartone might be influenced by variable extravascular factors suchas oxygentension.Therefore,examinationof vessels separate from pulmonary parenchyma might be the preferred method for evaluating direct vascular responses. In addition,vascular distendingpressurecould be maintained at a constant value;it did not vary in our study despite large changes in vessel dimension, which may occur during active vasomotion. The approach used minimized the manipulation of microvessels, especially the lumen. This decreases the likelihood of traumatic endothelial and smooth muscle damagecompared withthe manipulationrequired for the

Shafique et al. 4 8 5

measurement of tensionof microvascularrings and other approaches. Becauseof the small sizeof the vessels examined in the study, certain manipulationssuch as endothelialdenudation are difficultand, at times, unr~liable. However,metabolic blockingagents such as L-NMMA, which chemically remove the influence of endothelium-derived substances,may be used to ascertain the mechanisms of the vasoactive agents studied. Because experiments were performed in vitro, direct inferences cannot be made regarding the regulation of pulmonary perfusion. However, in most comparative studies, findings in vitro are similar to findings in vivo. Clinical implications. Because pulmonary hypertensionand right-sidedheart failure are not uncommonafter cardiac operationsin which CPB is used, altered vasomotor tone of pulmonary resistance vessels may have considerable impact on postoperativerecovery. The present study demonstrates altered endothelium-dependent pulmonary microvascular responses to acetylcholine and serotoninafter total CPB with associatedrelative pulmonary vascular ischemia and reperfusionafter cessationof CPB. The role of cholinergic, serotonergic, and other endothelium-dependent mechanisms in regulating ventilation and perfusion is not known. It is possible that altered vascular reactivity may disrupt normal pulmonary flow regulation and lead to ventilation-perfusion mismatch with associated shunting. This study shows an increased generation of a constrictor prostaglandin by pulmonary microvessels after CPB. Although the substance was not biochemically identified, it may be thromboxane A2,a constrictor prostaglandin, which is being implicated increasingly in anomaliesof pulmonary physiology.PAn increase in the systemic levels of thromboxane has been shown during clinicalCPB,24 but the clinicalrelevanceof this findingis speculative. Whatever the mechanism, the role of modified techniques in the maintenance of pulmonary perfusion during operation, such as partial rather than total CPB, or additivesto reduce pulmonary reperfusion injury deserve further investigation.

1.

REFERENCES Kato M,StaubNC. Response ofsmall pulmonary arteries to unilobular hypoxia and hypercapnia. Circ Res 1966; 19:426-40.

Ignarro LJ, Byrns RE,Wood KS. Endothelium-dependent modulation of cGMP levels and intrinsic smooth muscle tone in isolated bovineintrapulmonary arteryandvein. Circ Res 1987;60:82-92. 3. Chenworth DE, Cooper SW, Hugli TE, Stewart RW, Blackstone EH, Kirklin JW. Complement activation dur-

2.

4 8 6 Shafique et al.

ing cardiopulmonary bypass: evidence for generation of C3a and C5a anaphylatoxins. N EnglJ Med 1981 ;304:497503. 4. Mehta JL, Nicholas WW, Donnelley WH, Lawson DL, Saldeen TGP. Impaired canine coronary vasodilator responses to acetylcholine and bradykinin after occlusionreperfusion. Circ Res 1989;64:43-54. 5. Quillen JE, Sellke FW, BrooksLA, Harrison DG. Ischemia-reperfusion impairsendothelium-dependent relaxation of coronary microvessels but does not affect large arteries. Circulation 1990;82:586-94. 6. Cherry PD, GillisCN. Evidence for the role of endothelium-derived relaxingfactorinacetylcholine-induced vasodilatation in the intact lung. J Pharmacol Exp Ther 1987; 241:516-20. 7. Hayman AL, KadowitzPJ, Lippton HL. Methyleneblue selectively inhibits pulmonaryvasodilator responses in cat. J Appl Physiol1989;66:1513-7. 8. McMahon TJ, Hood JS, KadowitzPJ. Pulmonaryvasodilator response to vagal stimulationis blocked by Nw-nitroL-arginine methyl ester in the cat. Circ Res 1992;70:364-9. 9. Kadowitz PJ, Hayman AL. Influence of a prostaglandin endoperoxide analogue on the canine pulmonaryvascular bed. Circ Res 1977;40:282-7. 10. Hayman AL, KadowitzPJ. Pulmonary vasodilator activity of prostacyclin (PGh) in the cat. Circ Res 1979;45: 404-9. 11. Hayman AL, KadowitzPJ. Influence of tone on responses to acetylcholine in the rabbit pulmonary vascular bed. J Appl Physiol1989;67:1388-94. 12. Shirai M, Ninomiya I, Sada K. ThromboxaneAz/endoperoxidereceptorsmediatecholinergic constriction of rabbit lung microvessels. J Appl Physiol 1992;72:1179-85. 13. Shafique T, Sellke FW, Thurer RL, Weintraub RM, Johnson RG. Cardiopulmonary bypass and pulmonary thromboxanegeneration.Ann Thorac Surg 1993;55:724-8. 14. Biaggioni I, King LS, Enayat N, Robertson D, Newman

The Journal of Thoracic ·and Cardiovascular Surgery September 1993

JH. Adenosine produces pulmonary vasoconstriction in sheep: evidence for thromboxaneAz/prostaglandinendoperoxide-receptor activation. Circ Res 1989;65:1516-25. 15. JohnsRA, LindenJM, Peach MJ. Endothelium-dependent relaxationand cyclicGMP accumulation in rabbit pulmonary artery are selectively impaired by moderate hypoxia. Circ Res 1989;65:1508-15. 16. AdnotS, RaffestinB,EddahibiS, BraquetP, ChabrierpoE. Lossof endothelial-dependent relaxant activityin the pulmonary circulation of rats exposed to chronic hypoxia. J Clin Invest 1991;87:155-62. 17. Ku DD. Coronary vascular reactivityafter acute myocardial ischemia. Science 1982;218:576-8. 18. Ratych RE, Chuknyiska RS, Bulkley GB. The primary localization of free radical generation after anoxia/reoxygenation in isolated endothelial cells. Surgery 1987;102: 122-31. 19. Rubanyi GM, Vanhoutte PM. Superoxide anions and hyperoxia inactivate endothelium-derived relaxingfactor. Am J Physiol1986;250:H822-7. 20. Stewart DJ, Pohl D, Bassenge E. Free radicals inhibit endothelium-dependent dilationincoronaryresistance bed. Am J Physiol1988;255:H765-9. 21. RatliffNB, YoungWG, HackelDB,MikatE, WilsonJW. Pulmonaryinjury secondaryto extracorporeal circulation: an ultrastructural study. J THORAC CARDIOVASC SURG 1973;65:425-32. 22. HorpacsyG, Hugel W, Muller H, Geisler A. Biochemical monitoring of the lung during and after extracorporeal circulation. Prog Clin BioI Res 1987;236A:95-106. 23. Klausner JM, Paterson IS, Goldman G, et al. Thromboxane A2 mediates increasedpulmonarymicrovascular permeabilityfollowing limb ischemia. Circ Res 1989;64:117889. 24. Davis GC, Sobel M, Salzman EW. Elevated plasma fibrinopeptides A and thromboxane B2 levels during cardiopulmonary bypass. Circulation 1980;61 :808-14.