Role of Blood Components in Mediating Lung Vascular Injury after Pulmonary Vascular Thrombosis

Table 1-SUIIIDICif'Yf{.Raulla for Neutrophil-Deplelion Seudia in Babbila

Group 1 2 3 4

Neutrophils in Lung

Extravascular Lung Water

Nitrogen Mustard Rx

Inspired Gas

Cin:ulating Neutropbilslcmm

10 High Power Fields

Dry Lung llssue

+

o. o.

17±19 4422±2119 29±26 7633±2967

6±7 62±30 0 2±3

6.7± 1.5 6.7±1.3 3.5±0.7 3.7±0.2

+

Air Air

Survival Thne (hrs) 75±10 70±7

Numbers are mean ± 1 standard deviation.

64 ± 21 hours of neutropenia, bad no pulmonary edema. We conclude that neutrophil depletion does not prevent or reduce the severity of oxygen-induced lung injury in adult rabbits. ACKNOWLEDGMENT: Work presented in this paper was supported by U.S. Public Health Service Program Project Grant HL-25816. The work was accomplished during Dr. Bland·s tenure as an Established Investigator of the American Heart Association. REFERENCES

1 Craddock PR, Fehr J, Brigham K, Kronenberg K, Jacob HS. Complement and leukocyte-mediated pulmonary dysfunction in hemodialysis. N Engl J Med 1977; 296:769-74 2 Johnson A, Malik AB. Effect of granulocytopenia on extravascular lung water content after microembolization. Am Rev Respir

Dis 1980; 122:561-66

3 Heflin AC, Brigham KL. Prevention by granulocyte depletion of increased vascular permeability of sheep lung following endotoxemia. J Clin Invest 1981; 68:1253-60 4 Flick MR, Perel A, Staub NC. Leukocytes are required for increased lung microvascular permeability after microembolization in sheep. Circ Res 1981; 48:344-51 5 Fox RB, Hoidal JR. Brown DM, Repine JE. Pulmonary inflammation due to oxygen toxicity: involvement ofchemotactic factors and polymorphonuclear leukocytes. Am Rev Respir Dis 1981; 123:521-23 6 Shasby DM, Fox RB, Harada RN, Repine JE. Mechanisms of pulmonary oxygen toxicity: neutropenia protects against acute lung injury from hyperoxia. J Appl Phys 1982; 52:1237-44 7 Martin WJ U, Gadek JE, Hunningbake GW, Crystal RG. Oxidant injury of lung parenchymal cells. J Clin Invest 1981; 68:1277-88 8 Bowman CM, Harada RN, DeLong S, Vatter AE, Repine JE. Hyperoxia damages endothelial cells in tissue culture. Pediatr Res 1981; 15:715 (abstract) 9 Pearce ML, Yamashita J, Beazell J. Measurement of pulmonary edema. Circ Res 1965; 16:482-88 10 Erdmann AJ m. Vaughan TR. Jr. Brigham KL. Woolverton we, Staub NC. Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ Res 1975; 37:271-284

Role of Blood Components In Mediating Lung vascular Injury after Pulmonary vascular ThrombOsis* A B. Malik, Ph.D.; A ]ohfl80fl, Ph.D.; M. V. Tahamont, Ph.D.; Hoyte tl611 tier Zee, M.D.; and F. A Blumenstock, Ph.D. *From the Department of Physiology, Albany Medical College of Union University, Albany. Supported ~y grants HL-17355, HL-26551, HL-07529 (NRSA Post Doctoral Fellowship to A. Johnson), and Parker B. Francis Foundation Fellowship to M. V. 1abamont. Reprint ntQ!188tB: Dr. Malik, Physiology, Room ME331, Alba1111 Medical College, 47 New Scotland Avenue, Albanyl2208.

Dulmonary microembolization is a characteristic feature of .C the adult respiratory distress syndrome. 1 The microemboli consist of fibrin clots and of leukocytes and platelets trapped in small pulmonary vessels. Previous studies have characterized the effects of pulmonary microembolization on lung ftuid and protein exchange. It is clear now that embolization by whatever means increases the pulmonary vascular permeability to proteins indicating that pulmonary microembolization leads to lung vascular injury. 1 This phenomenon is observed in both sheep and dog lymph preparations. a• In this report, we summarize the effects of thrombininduced pulmonary microembolization on lung ftuid and protein exchange and our results dealing with the mechanisms by which these alterations are induced. Thrombin was chosen as a method for inducing pulmonary microembolization because it converts fibrinogen directly to fibrin, which in tum results in the activation of plasmin, and finally the activation of the complement system. 5 The lung vascular injury after pulmonary microembolization is the complex effect ofactivation of these systems. • In these studies we have examined the independent effects of some of the key bloodborne factors in the mediation oflung vascular injury. Studies were made in intact sheep in which pulmonary lymph was collected. PLATELETS

We examined the effect of platelets in mediating the increase in lung vascular permeability after pulmonary microembolization by depleting circulating platelets. The platelet count was decreased by 97% of its baseline value in sheep using anti-platelet serum prepared against sheep platelets. Despite the thrombocytopenia, pulmonary lymph ftow increased after thrombin infusion and there was no change in the lymph-to-plasma protein concentration ratio ('Thble 1). This response was similar to that observed in the control group after thrombin ('Thble 1). Platelet depletion did not prevent the thrombin-induced increase in lung vascular permeability since raising the pulmonary microvascular pressure (Pmv) by inftation of a left atrial balloon produced a relatively large increase in lymph ftow without a change in the lymph-to-plasma protein concentration ratio; this response was the same as observed in control thrombin group after the increase in Pmv. Therefore, platelet depletion per se did not prevent the thrombin-induced increase in lung vascular permeability, suggesting that platelet aggregation did not mediate the response. GRANULOCYTES

The role of granulocytes was examined by selectively depleting the circulating granulocyte levels with hydroxyurea (200 mglkg per day for 4 days). Hydroxyurea did not CHEST I 83 I 5 I -

1883 I Supplement

218

Table 1-Stetulr/ State Claangea Ajt.r Thrombin-Induced M~ Group Thrombin Control (n=8) Baseline Post-Thrombin Granulocyte Depletion (n=7) Baseline Post-Thrombin Platelet Depletion (n=5) Baseline Post-Thrombin

Qlym (mllhr)

UP

Protein Clearance (mllhr)

(mm Hg)

(mm Hg)

Pr.

QL (Umin)

PVR (mm Hg/Umin)

6.89±1.63 13.53±2.81*

0.79±0.04 0.82±0.04

5.51±1.46 11.26±2.52*

15.9±2.09 29.9±2.58*

3.25±0.50 3.19±0.53

1.98±0.16 1.39±0.25

6.45±1.19 17.42±2.46*

5.81±1.53 6.67±1.30

0.88±0.04 0.79±0.06

5.43±1.5 5.64±1.32

12.4±1.29 22.9±3.12*

4.71±0.84 5.21±0.92

1.87±0.29 1.67±0.20

4.51±0.77 11.66±2.50*

7.23±1.18 11.42± 1. 75*

0.77±0.06 0.75±0.05

5.62±0.92 8.53±1.31*

15.4±1.63 35.2±4.87*

4.90±1.80 5.40±1.43

1.89±0.16 2.13±0.81

5.72±0.82 17.29±3.06*

Pp.

Qlym = pulmonary lymph flow; UP ratio = lymph-to-plasma protein concentration ratio; protein clearance = Qlym x UP ratio; Pp. = mean pulmonary arterial pressure; Pi; = mean left atrial pressure; Q. = pulmonary blood flow; PVR = pulmonary vascular resistance; *p<0.05 from baseline.

significantly affect the lymphocyte count while it only reduced the granulocyte count by 98% of pretreatment value. Thrombin infusion produced a 3-fold increase in the vascular resistance but without significant changes in pulmonary lymph flow or lymph-to-plasma protein concentration ratio (Table I). Granulocyte depletion prevented the increase in fluid filtration that characteristically occurs after thrombininduced pulmonary microembolization. When the left atrial pressure was increased to increase Pmv, pulmonary lymph flow increased but the increase in lymph flow was associated with a decrease in lymph-to-plasma protein concentration ratio. The response was similar to that occurring in normal lungs after an increase in Pmv. Therefore, granulocyte depletion prevented the thrombin-induced increase in lung vascular permeability, suggesting that granulocytes mediate the increased permeability. FIBRIN

in defibrinogenated sheep produced relatively small increases in pulmonary artery pressure and pulmonary vascular resistance as compared to the control group, probably because thrombin did not induce formation of fibrin clots in the defibrinogenated animals.• Also, thrombin infusion produced a small increase in the pulmonary lymph flow which was associated with a decrease in lymph-to-plasma protein concentration ratio. Raising Pmv in this group with a left atrial balloon produced further increase in lymph flow, which was associated with a further decrease in the lymph-toplasma protein concentration ratio indicating that defibrinogenation prevented the thrombin-induced increase in lung vascular permeability. Therefore, formation of fibrin clots is necessary for the production of lung vascular injury since thrombin did not increase lung vascular permeability without fibrinogen. Intravascular fibrin is a requirement for the development of thrombin-induced lung vascular injury.

Studies were made to determine ifthe thrombin-induced FIBRINOLYTIC PRooucrs increase in lung vascular permeability is dependent on the presence of circulating fibrinogen. Ifthis is the case, then Since intravascular fibrin is required for the development fibrinogen depletion with Ancrod should also prevent the of lung vascular injury after thrombin, SaldeenL• has proincrease in lung vascular permeability. On the other hand, posed that activation of fibrinolysis, by conversion of plasminogen to plasmin, is the inciting event which results in thrombin may have a direct effect that is independent of its effect in converting fibrinogen to fibrin; for example, thromcomplement activation, and this in tum leads to the generation of chemotactic fragments such as C5a which produce bin directly causes platelet aggregation which may in tum pulmonary leukostasis. 5 ' 7 The activation of fibrinolysis is increase transvascular fluid filtration in the lung. Fibrinogen was depleted with chronic treatment with Ancrod (a purified important since it generates not only plasmin, but also fibrin degradation products which are themselves chemotactic. fraction of the Malayan pit viper venom). Thrombin infusion Table 1-Ejfect. rf Defibrinogetllllion tmd Fibrinolgaia Inlaibition Qlym (mllhr)

UP

Protein Clearance (mllhr)

(mm Hg)

(mm Hg)

Pr.

QL (Umin)

PVR (mm Hg/Umin)

Defibrinogenation (n=6) Baseline Post-Thrombin

8.7±2.7 11.6±3.1*

0.78±0.06 0.70±0.05*

6.6±1.9 8.1± 1.9*

13.8±1.8 18.0±2.9*

3.3±1.0 4.2±1.0

1.8±0.2 1.7±0.3

6.1±1.1 9.3±2.1

Fibrinolysis Inhibition (n=l3) Baseline Post-Thrombin

5.4±1.2 11.1± 1.4*

0.74±0.03 0.64±0.03*

4.2±1.1 6.6±0.9*

18.8±1.1 30.3±1.5*

5.5±0.6 7.6±0.8

2.1±0.2 1.3±0.2*

6.9±0.9 23.1±42*

Group

Pp.

*different from baseline 228

Lung Delanu,

I~

and Repair

'Dible 3-Effeet tfCornpkrnent Depletion on f1w .Raponae to Thrombin (n=4)

Qlym Baseline Post-Thrombin

(mllhr)

UP

5.7±1.9 6.7±2.0

0.73±0.06 0.61±0.07

Pr.

QL

Protein Clearance

Pp;-

(mllhr)

(mm Hg)

(mm Hg)

(Umin)

(mm HWL'min)

4.4±1.8 4.0±1.4

16.3±2.2 20.8±3.5*

4.3±1.6 4.3±1.7

2.6±0.4 1.8±0.6

5.2±1.1 11.9±5.2*

PVR

*p<0.05 from baseline

Because of the pivotal role of fibrinolysis activation, we reasoned that complete inhibition of fibrinolysis with tranexamic acid, which would inhibit activation of plasmin and generation of fibrin degradation products, should have a protective eBect ifSaldeen's hypothesis is coJTeCt. Thrombin infusion normally generates fibrin degradation products indicating the activation of plasmin bas occurred; however, in animals pretreated with tranexamic acid, concentration of fibrin degradation products did not increase after thromin infusion. In this group, thrombin infusion produced increase in pulmonary lymph ftow which was associated with a decrease in the lymph-to-plasma protein concentration ratio ('Ikble 2). Raising Pmv in this group produced a further increase in lymph ftow and decrease in lymph-to-plasma protein concentration ratio indicating inhibition of fibrinolysis prevented the increase of lung vascular pel'meability. TherefOre, activation of fibrinolysis is another requirement for the production of thrombin-induced increase in lung vascular permeability.

activation may be necessary for the subsequent granulocyte activation. Complement activation likely occurs via the generation of plasmin rather than thrombin per se because thrombin infusion in defibrinogenated sheep did not increase lung vascular permeability (Thhle 2). SUMMARY

Figure 1 indicates the postulated mechanisms of the increased lung vascular permeability after thrombin-induced pulmonary microembolization. Thrombin converts fibrinogen to fibrin, and fibrin is a necessary requirement for the mediation oflung vascular injury because thrombin failed to increase lung vascular permeability in the absence of fibrinogen (induced by Ancrod). Fibrin may act by 2 mechanisms: (1) by inducing the activation of plasmin, and (2) by serving as a source of 6brin degradation products. Plasmin activation is important since inhibiting plasminogen activation with tranexamic acid also prevented the increase in lung vascular permeability. Plasmin may act by activating the complement system to generate chemotactic fragments such as C5a and also by lysing 6brin to generate fibrin degradation products which are also chemotactic. Granulocyte aggregation is a crucial step in the genesis of lung vascular injury because granulocyte depletion with hydroxyurea prevented the increased pulmonary vascular permeability. The final step is not certain, but may be related to generation of oxygen radicals and proteases from the activated granulocytes. Platelets do not 6gure into this scheme because platelet depletion failed to prevent the increase in lung vascular permeability after activation of the clotting mechanism with thrombin.

COMPLEMENT PROTEINS

Since granulocyte depletion prevented the increase in lung vascular permeability after thrombin, the question arises whether there is activation of the complement system which in turn causes granulocyte sequestration in pulmonary microvessels. Complement activation can occur secondary to plasmin activation, 5 and thus is linked to formation of intravascular fibrin clots. Iflung vascular injury is dependent on the complement activation, then prior complement depletion with the cobra venom fBctor should be as protective as granulocyte depletion. Thrombin infusion in sheep, in which complement was depleted with cobra venom fBctor, did not change the pulmonary lymph ftow and the lymph-toplasma protein concentration ratio (Thble 3). The increase in Pmv induced by a left atrial balloon catheter increased lymph ftow, but decreased the lymph-to-plasma protein concentration ratio. TherefOre, the complement system is required for the production of thrombin-induced lung vascular injury suggesting that it is necessary to activate the complement system to increase lung vascular permeability. Complement

~m.grc

F""

Pl..AsMJt«lGEN COt>I..MNT I

I!:WI

Ill

/'~

j'"

REFERENCES 1 Saldeen T. The microembolism syndrome. In: Saldeen "I: ed, The microembolism syndrome. Stockholm: Almquist and Wlksell, 1979; 7-44 2 Obkuda K, Nalcabara K, Weidner WJ, Binder AS, Staub NC. Lung fluid exchange after uneven pulmonary artery obstruction in

sheep. Circ Res 1978; 53:152-61 3 SaldeenT. Fibrin-derivedpeptides. In: MalikAB, StaubNC, eds,

A1"':-ml

""RP.

._ .

-...Giwu..ocvrE

t

CSA,-----.- AGGREGATI<*

Q •

?'---..

IJ.t«; VASCWR

INJl.RY

111111

FIGURE 1. Postulated mechanisms of lung vascular injury after thrombin-induced pulmonary microembolization. CVF = cobra venom factor; FDP = fibrin degradation products; HDY = hydroxyurea. CHEST I 83 I 5 I -

1883 I Supplement

238

4 5

6

7

Mechanisms of lung microvascular injury. Ann New York Acad Sci 1982; 384:319-31 Malik AB, van der Zee H. Lung vascular permeability fOllowing progressive pulmonary microembolization. J Appl Physiol 1978; 45:590-97 Kaplan AP, Silverberg M, Dunn JT, Ghebrehiwet B. In: Kushner I, Volankis JE, Gewurz H. eds, C-reactive protein and the plasma protein response to tissue injury. Ann New York Acad Sci 1982; 389:25-38 Malik AB, Johnson A, 18hamont MV. Mechanisms oflung vascular injury after intravascular coagulation. In: Malik AB, Staub NC, eds, Mechanism of lung vascular injury. Ann New York Acad Sci 1982; 384:213-34 Jacobs HS, Craddock PR, Hammerschmidt DE, Moldow CF. Complement-induced granulocyte aggregation. N Eng) J Med 1980; 302:789-74

Pulmonary Vasoconstriction and Profound Leukopenia In 1\vo Sheep Experimental Models* Effects of Complement

leukopenia, we depleted sheep of complement with cobra venom factor prior to either bypass or endotoxin infusion. METHODS

Experiments were carried out in 13 awake SufiOlk sheep weighing 24-34 kg. Intravascular catheterization allowed continuous measurements ofpulmonary and systemic vascular pressures. Cardiac output was measured by thermodilution. ECMO was perfOrmed by partial vena-venous bypass ill heparinized sheep using a new 0.8 M2 spiral coil Silicone membrane lung and an occlusive roller pump. 3 The membrane lung was ventilated with I(J()IJ, oxygen and pump ftow averaged 750 ml/min. E coli endotoxin (LPS W. Oill:B4,Difco) 1JLg/kg was intravenously infused over 5 min. Sheep were depleted of complement over 3 days by serial intravenous injections of cobra venom factor (10 Ulkg x 7, Naje Haje, Cordis, Miami, FL). Total hemolytic sheep complement (CH50) levels were measured using rabbit red cells sensitized with goat anti-rabbit red cell antibody. Sheep C3 levels were measured by a radial immunodiffusion assay using rabbit anti-sheep C3 antibody (Cappel Lab, Cochranville, PA). Plasma TxB. was assayed with a double antibody radioimmunoassay technique.• Leukocytes were counted with a Coulter cell counter and platelets by phase microscopy.

Depletion

RESUI.J'S AND DISCUSSIONS

P. C. Hiittemeier, M.D.; D. BerT'fl, M.D.; K. J Bloch, M.D.; W. D. Watkina, M.D.; and W. M. Zapol, M.D., F.C.C.P.

Endotoxin infusion or bypass in nonnal sheep decreased

In the awake sheep, the acute and transient increase of pulmonary vascular resistance fOllowing either venovenous cardiopulmonary bypass or E coli endotoxin infusion is primarily mediated by pulmonary synthesis of thromboxane. l-3 Inhibition of pulmonary vasoconstriction by pretreatment with either cyclooxygenase or thromboxane synthetase inhibitors does not prevent the leukopenia resulting from either bypass or endotoxin. l-3 Thus, certain metabolic similarities exist between these two models, although their fundamental mechanisms probably differ. Intravascular activation of complement may be central to both models since: 1) intravenous infusion of autologous plasma in which complement was activated with zymosan has been reported to produce acute eicosanoid-mediated pulmonary vasoconstriction and leukopenia in the awake sheep;• 2) both bacterial endotoxins and exposing blood to polymer surfaces can activate plasma complement;5 '8 3) CS fragments resulting from complement activation induce polymorphonuclear leukocyte aggregation in vitro. 7 To compare the mechanisms ofarachidonate activation and *From the Departments of Anesthesia and Medicine, Massachusetts General Hospital, Harvard Medical School, Boston. Supported by NHLBI (SCOR) grant HL-23591 in Adult Respiratory Failure. &print re~: Dr. Zapol, Depamnent of Anesthuio, Mt1880Chusett& General HO&pital, Boston onl4

CH50 (25%) and C3 (30%), suggesting complement was activated. Administration of cobra venom factor reduced CH50 by >95% and C3 by >90%. Thble 1 summarizes the results of complement depletion in bypass and endotoxintreated sheep. Complement depletion prevented an increase of pulmonary artery pressure during bypass, but did not affect endotoxin-induced pulmonary artery hypertension. Thromboxane production during bypass requires an intact complement system, while increases of plasma Tx82 fOllowing endotoxin infusion were uninhibited by complement depletion. The transient leukopenia of bypass appears to be complement mediated, since it was absent during bypass of complement depleted sheep. Endotoxin-induced leukopenia occurred to the same extent as in nonnal sheep despite complement depletion. These results highlight a fundamental difference with respect to the activation of arachidonate metabolism in these two experimental models. As shown in Thble 1, several other basic differences exist between these 2 models in their nonnal state. The maximum levels of plasma TxB2 and pulmonary artery pressure are much greater fOllowing endotoxin than during bypass. The leukopenia of bypass is rapid and transient lasting only 45-60 min while it is slower in onset after endotoxin and irreversible. Acute and irreversible thrombocytopenia occurs during bypass, but is not observed after endotoxin infusion. The early platelet loss fOllowing the onset of bypass may partially

Table 1

VV Bypass

Stimulus

Response Normal sheep Complement depleted

mean peak PAP (mm Hg) 31 17

mean peak Tx81 (nglml) 1.2 <0.1

Endotoxin

Leukopenia

Thrombocytopenia

++

(transient) 0

+

mean peak PAP (mm Hg)

TxB1 (nglml)

Leukopenia

Thrombocytopenia

40

9.7

+

0

38

11.1

+

0

mean peak

+present; PAP= pulmonary artery pressure; 0 absent 248

Lung Deflnle,

~ury

and Repelr