5G promoter polymorphism in the plasminogen-activator-inhibitor-1 gene and outcome of meningococcal disease

Early reports

4G/5G promoter polymorphism in the plasminogen-activatorinhibitor-1 gene and outcome of meningococcal disease Peter W M Hermans, Martin L Hibberd, Robert Booy, Olufunmilayo Daramola, Jan A Hazelzet, Ronald de Groot, Michael Levin, and the Meningococcal Research Group*

Summary Background Intravascular coagulation with infarction of skin, digits, and limbs is a characteristic feature of meningococcal sepsis. Children with meningococcal sepsis have higher than normal concentrations of plasminogen activator inhibitor 1 (PAI-1) in plasma. Combined with the widespread venous thrombosis, this finding suggests an impairment of fibrinolysis. A common functional insertion/deletion (4G/5G) polymorphism exists in the promoter region of the PAI-1 gene. We tested the hypothesis that children with the 4G/4G genotype produce higher concentrations of PAI-1, develop more severe coagulopathy, and are at greater risk of death during meningococcal sepsis. Methods The relation between meningococcal disease outcome, PAI-1 concentration, and PAI-1 genotype was investigated in 175 children with meningococcal disease (37 from Rotterdam, the Netherlands, and 138 from London, UK) and 226 controls (137 from Rotterdam, 89 from London). PAI-1 concentrations in plasma were measured by ELISA, and the 4G/5G PAI-1 polymorphism was detected by PCR and hybridisation. Findings Concentrations of PAI-1 on admission correlated with presentation (sepsis or meningitis) and outcome. The median PAI-1 concentration in children who died was substantially higher than that in survivors (2448 [IQR 1115–3191] vs 370 [146–914] ng/mL; p<0·0001). Patients with the 4G/4G genotype had significantly higher PAI-1 concentrations than those with the 4G/5G or 5G/5G genotype (1051 [550–2440] vs 436 [198–1225] ng/mL; p=0·03), and had an increased risk of death (relative risk 2·0 [1·0–3·8] for the two cohorts combined, and 4·8 [1·8–13] for the London cohort). Interpretation A genetic predisposition to produce high concentrations of PAI-1 is associated with poor outcome of meningococcal sepsis. This finding suggests that impaired fibrinolysis is an important factor in the pathophysiology of meningococcal sepsis. Lancet 1999; 354: 556–60 *Investigators given at the end of the paper Department of Paediatrics, Sophia Children’s Hospital, Erasmus University Rotterdam, Rotterdam, Netherlands (P W M Hermans PhD, J A Hazelzet MD, Prof R de Groot MD); Department of Paediatrics, Imperial College School of Medicine at St Mary’s Hospital (M L Hibberd PhD, O Daramola MSc, Prof M Levin FRCP ); and Department of Epidemiology and Public Health, Institute of Child Health, London, UK (R Booy FRCAP) Correspondence to: Dr Peter W M Hermans, Laboratory of Paediatrics, Erasmus University Rotterdam, PO Box 1738, 3000 DR Rotterdam, Netherlands (e-mail: [email protected])

Introduction Meningococcal disease remains an important cause of childhood mortality and morbidity in both more developed and less developed countries.1 Although meningococci are readily killed by antibiotics, the organisms or meningococcal products (mainly endotoxin) trigger an intense host inflammatory response which leads to widespread endothelial injury, with tissue damage and organ failure.2 The severity of meningococcal disease is related to the concentration of endotoxin,3 which in turn is related to higher concentrations of inflammatory mediators such as tumour necrosis factor ␣ and interleukin-1␤.4 Patients with fulminant meningococcal sepsis have evidence of activation of neutrophils,5 monocytes,4,6 and both the coagulation7 and complement cascades.8 Coagulopathy is a consistent feature of meningococcal sepsis, and is characterised by longer than normal prothrombin and partial thromboplastin times, raised concentrations of fibrin degradation products, depletion of coagulation factors, and thrombocytopenia.7,9,10 Virtually all patients show some evidence of coagulation derangement, but abnormalities are most marked in those with severe septic shock.2,7,9,10 In patients with the clinical syndrome of purpura fulminans, there is severe disseminated intravascular coagulation with widespread thrombosis, resulting in necrosis of large areas of the skin and infarction of limbs or digits.2,7,9 Many different approaches have been suggested for the treatment of the coagulopathy of meningococcal sepsis and purpura fulminans.11 These have included the use of heparin,12 prostacyclin,13 infusions of concentrates of protein C,14 antithrombin III,15 tissue plasminogen activator,16 and plasmapheresis17 or haemofiltration14 to remove inflammatory mediators. Unfortunately, the lack of a clear understanding of the mechanisms leading to the coagulopathy has made selection of the most appropriate form of treatment difficult. The pathophysiology of the coagulopathy of meningococcal sepsis seems to involve an imbalance of procoagulant and anticoagulant mechanisms.2,9,10 Concentrations of the anticoagulant proteins C, S, and antithrombin III in plasma are low in meningococcal sepsis,7,10 and in-vitro endothelial production of prostacyclin is impaired by exposure to plasma from meningococcal patients.13 Thrombomodulin may be down-regulated on the endothelial surface,18 and procoagulant molecules such as tissue factor are increased on monocytes6, leading to fibrin deposition and formation of microthrombi. Widespread venous thrombosis also suggests impairment of fibrinolytic pathways.19 Endotoxin infusion in healthy individuals induces up-regulation of both tissue plasminogen activator and plasminogen activator

Group

All patients n

Fatal cases Median (IQR) PAI-1 (ng/mL)

Non-fatal cases

n

Median (IQR) PAI-1 (ng/mL)

n

Median (IQR) PAI-1 (ng/mL)

All cases

123

472 (164–1387)

21

2448 (1115–3191)

102

370 (146–914)

Sepsis All sepsis Rotterdam London

93 36 57

759 (258–1996) 1050 (494–2710) 475 (201–1261)

18 11 7

2460 (1705–3370) 3155 (1527–3602) 2440 (1889–3110)

75 25 50

555 (206–1173) 686 (366–1753) 431 (167–1036)

Mixed disease

19

190 (67–765)

1

1002*

18

189 (47–410)

Meningitis

11

50 (31–174)

2

9

45 (27–174)

45, 59*

*Actual values since numbers were small.

Table 1: Relation between PAI-1 concentrations, clinical presentation, and disease outcome

inhibitor (PAI-1).20 PAI-1 is one of the serpin protease inhibitors,21 which forms a one/one complex with its target proteolytic enzyme. PAI-1 reacts with both tissue and urinary plasminogen activators, as well as with protein C. 22 It is synthesised and secreted by various cell types including endothelial cells, hepatocytes, and platelets.21,23 This protein is an acute-phase reactant, and its concentrations are increased by inflammatory stimuli such as interleukin-1␤ and tumour necrosis factor ␣.23,24 Concentrations of PAI-1 in plasma are very high in children with meningococcal sepsis, the highest concentrations being found in severe and fatal disease.25 A common functional polymorphism exists in the PAI1 gene. A single base-pair insertion (5G)/deletion (4G) polymorphism 675 bp upstream from the start of transcription is functionally important in regulating the expression of the PAI-1 gene.24 Artificial constructs containing the PAI-1 promoter have shown that the 4G allele produces six times more RNA than the 5G allele in response to interleukin-1␤.24 Individuals homozygous for the 4G allele have higher basal and inducible concentrations of PAI-1 than those with one or two copies of the 5G allele.24 We postulated that compared with children without the polymorphism, those with the 4G/4G genotype would produce higher concentrations of PAI-1, develop a more severe coagulopathy, and have a greater risk of death if they acquired meningococcal sepsis. We now report a study on a large cohort of patients with meningococcal disease.

Patients and methods Patients Between 1991 and 1995, the paediatric intensive-care units at St Mary’s Hospital, London, UK, and at Sophia Children’s Hospital, Rotterdam, Netherlands, participated in a multicentre study of antiendotoxin therapy in meningococcal disease. This study established uniform diagnostic criteria and clinical management protocols for the patients in both centres. External monitoring of the trial centres confirmed that the management and clinical outcome were similar.26 This confirmation, together with the known similarity in ethnic origin of the population (largely white Europeans), led us to undertake a combined study of patients and controls from both centres. Meningococcal disease was diagnosed in patients presenting with characteristic purpuric rash, fever, and meningitis, septicaemia, or both. Patients were judged to have meningococcal meningitis if they had evidence of meningism without the presence of shock or impaired peripheral perfusion, and meningococcal sepsis if they presented without evidence of meningism and with features of shock or impaired peripheral perfusion (prolonged capillary refill time, tachycardia, raised base deficit, oliguria, and impaired oxygenation). Patients with evidence of both shock and meningitis were classified as having mixed disease. Diagnosis was confirmed by isolation of meningococci from blood or cerebrospinal fluid, detection of meningococcal antigens, or by PCR detection of meningococcal

genome in blood or cerebrospinal fluid; some cases were confirmed by more than one technique. The Rotterdam cohort comprised 37 children with meningococcal sepsis admitted to the paediatric intensive-care unit at the Sophia Children’s Hospital between 1991 and 1995. There were 21 boys and 16 girls. The mean age was 4·1 years (range 0·5–17·9). The London cohort consisted of 138 children with meningococcal disease admitted to the paediatric intensivecare unit at St Mary’s Hospital, between 1991 and 1995. There were 76 boys and 62 girls. The mean age was 3·5 years (range 0·3–17·9). Blood samples were obtained from 137 healthy, white Dutch infants who had been enrolled in a vaccine trial, and 89 healthy, white UK children whose families were friends of a family affected by meningococcal disease. The study was approved by the ethics committees of both hospitals, and blood samples were taken from patients and controls with informed parental consent.

Methods From the London and Rotterdam cohorts, 87 and 36 serum samples, respectively, were taken on admission to the intensivecare unit. Concentrations of PAI-1 antigen were measured in these samples by means of a commercially available ELISA kit (American Diagnostica, Alpha Laboratories Ltd, Eastleigh, UK). DNA was prepared from serum or blood by means of a standard small-scale isolation technique. An 890 bp region of the PAI-1 promoter from each sample was amplified by PCR, and the genotype was determined by allele-specific oligo melting.24 Controls for this technique consisted of DNA samples of known genotype determined by DNA sequence analysis.

Statistical analysis Since PAI-1 concentrations were not normally distributed, and because the variances of groups being compared were dissimilar, results are expressed as median (IQR) and were compared by

Figure 1: Concentrations of PAI-1 in plasma of children with different clinical presentations of meningococcal disease Points=median; bars=IQR.

Group

PAI-1 genotype

Patients

4G/4G

4G/5G

5G/5G

London

n

Median (IQR) PAI-1 (ng/mL)

n

Median (IQR) PAI-1 (ng/mL)

n

Median (IQR) PAI-1 (ng/mL)

Sepsis

All cases

33

1051 (400–2440)

67

374 (164–986)

23

467 (83–1225)

Sepsis All sepsis Rotterdam London

26 11 15

1487 (759–2440) 2125 (686–2511) 1085 (759–2414)

49 23 26

628 (280–1527) 986 (366–2801) 397 (201–728)

18 2 16

470 (78–1225) 551, 1115* 390 (75–1243)

Table 2: Relation between PAI-1 concentrations, genotype, and clinical presentation

PAI-1

means of the Kruskal-Wallis test. Genotype frequencies were compared by means of the ␹2 test (two-tailed) or Fisher’s exact test (two-tailed) when numbers were small.

Results As found in previous studies, the concentrations of PAI-1 on admission were substantially higher in the children with meningococcal disease (table 1) than the normal reference range for healthy controls (4–43 ng/mL). The concentrations were significantly higher in patients presenting with sepsis than among those with mixed disease or meningitis (p<0·0001). Among the children with sepsis, PAI-1 concentrations were higher in those from Rotterdam than in those from London (table 1). PAI-1 concentrations were higher in children who subsequently died than in those who survived, both in the whole group (p<0·0001) and in the sepsis subgroup (p<0·0001). No significant difference between the children who died and survivors was apparent in the mixed disease and meningitis subgroups, but numbers were small (figure 1). PAI-1 concentrations were significantly greater in the 33 patients with the 4G/4G genotype (1051 ng/mL [400–2440]) than in the 90 patients with the 4G/5G or 5G/5G genotypes (436 ng/mL [198–1225]; p=0·03). Further analysis of PAI-1 concentrations in relation to genotype showed that the higher concentration associated

Mixed disease

Meningitis Total

PAI-1 genotype 4G/4G 25 (27%) 4 (16%) 6 (29%) 4G/5G 46 (50%) 15 (60%) 10 (48%) 5G/5G 21 (23%) 6 (24%) 5 (24%) Total

*Actual values since numbers were small.

Controls Rotterdam * London

92

25

21

Rotterdam

Sepsis

35 (25%) 11 (30%) 71 (51%) 24 (65%) 32 (23%) 2 (5%) 138

37

24 (27%) 44 (49%) 21 (24%) 89

35 (26%) 70 (51%) 32 (23%) 137

*Only sepsis cases studied.

Table 3: PAI-1 promoter region genotype frequencies

with the 4G/4G genotype was significant for patients presenting with sepsis (table 2, figure 2; p=0·03), but not for the mixed or meningitis groups. This association was seen in both the Rotterdam cohort and the London sepsis cohort (table 2). The frequencies of the 4G/4G, 5G/4G and 5G/5G PAI-1 genotypes did not differ significantly between the combined patient cohorts and the combined control populations (27 vs 26%, 54 vs 50% and 19 vs 23%, respectively; p=0·7). Among the groups of patients, the genotype frequency did not vary significantly with respect to clinical diagnosis (sepsis, mixed, or meningitis; p=0·3; table 3). In the two cohorts combined, there was a significantly higher proportion of deaths among patients with the 4G/4G genotype (12 of 46 [26%]) than among those with the 4G/5G or 5G/5G genotypes (17 of 129 [13%]), giving a relative risk of 2·0 (95% CI 1·0–3·8). When analysed by diagnosis, the increased risk was apparent only in the sepsis subgroup (2·0 [1·0–4·0]), and not in the mixeddisease or meningitis subgroups. No significant association was seen when the smaller Rotterdam cohort was analysed independently (0·53 [0·1–2·1]; p=0·44), whereas the increased risk of death associated with the 4G/4G genotype was highly significant in the larger London cohort (4·8 [1·8–13·0]; p=0·002, table 4).

Discussion This study found a significant association between carriage of the homozygous 4G deletion polymorphism in the PAI-1 gene and mortality from meningococcal sepsis. The association is supported by the finding that concentrations of PAI-1 in plasma were higher in patients with the 4G/4G genotype than in patients with the 4G/5G or 5G/5G genotypes. PAI-1 concentrations were also related to the clinical presentation: patients with sepsis alone had significantly higher concentrations than those with meningitis alone, or those with mixed disease. Consistent with findings in previous studies, PAI-1 concentrations were strongly related to the severity of the disease, the highest values being found in patients who died. The results suggest that patients with the 4G/4G genotype respond to meningococcal sepsis by producing London Fatal PAI-1 genotype 4G/4G 4G/5G or 5G/5G Total

Figure 2: Relation between concentrations of PAI-1 in plasma, and PAI genotype in children with meningococcal sepsis Points=median; bars=IQR.

Non-fatal

Rotterdam

Combined

Fatal

Non-fatal

Fatal

Non-fatal

9 5

16 62

2 9

9 17

11 14

25 79

14

78

11

26

25

104

Table 4: Outcome of meningococcal sepsis in children classified by PAI-1 genotype

greater amounts of PAI-1. Since death is associated with higher PAI-1 concentrations, carriage of the 4G/4G genotype would be associated with a worse prognosis, and indeed our study has found this to be the case. Many studies have shown that patients presenting with meningitis have a different clinical picture from those presenting with meningococcal sepsis.2,8,10 Mortality is lower in those with meningitis, and the presence or absence of meningitis is a recognised component of several severity scores. The difference in PAI-1 concentrations between patients with meningitis and those with sepsis shown in this study highlights the fact that these two presentations of meningococcal disease probably differ in their pathophysiology. To establish an association between this functional genetic polymorphism and outcome of meningococcal disease, a large number of patients must be studied. Host factors such as the 4G/4G deletion polymorphism are likely to be some of many important determinants of outcome which might also include microbial differences and environmental effects. Our decision to combine patients from two different countries was based on the fact that these countries have mainly white populations, a similar incidence of meningococcal disease, and similar standards of medical care. The two centres from which patients were recruited (the Sophia Children’s Hospital, Rotterdam, and St Mary’s Hospital, London) have uniform clinical management procedures for children with meningococcal disease, and in particular took part in an externally monitored multicentre trial of an antiendotoxin antibody requiring similar diagnostic and clinical management procedures.26 Our findings that the UK and Dutch control children had similar frequencies of the PAI-1 genotypes supports our decision to combine the groups of patients, as does the fact that patients with the 4G/4G genotype had higher PAI-1 concentrations in both cohorts. The association between PAI-1 genotype and outcome, which was found in the combined group and in the larger London group, was not seen in the 37 patients from Rotterdam, although the confidence limits are overlapping. The higher case-fatality rate in the Rotterdam patients (table 1) is likely to be due to greater severity of disease in the Rotterdam patients than in the London patients, which is consistent with the PAI-1 concentrations being higher in Rotterdam patients. For a highly variable disease such as meningococcal disease, variation in results between subgroups is not surprising. No significant difference in the frequency of the three genotypes was found between patients with meningococcal disease and the control population. This finding suggests that the 4G/4G polymorphism is not associated with an increased susceptibility to meningococcal disease. The finding that a polymorphism that determines the expression of an inhibitor of the fibrinolytic pathway is significantly related to the outcome of meningococcal disease may seem surprising. However, there is increasing evidence that activation of inflammation and coagulation are closely related and mutually interdependent.27 In primitive species, there is a single pathway for both inflammation and coagulation.27,28 Activation of coagulation pathways is detected in virtually all patients with severe sepsis. Infusion of endotoxin into healthy volunteers is associated with an increase in the concentrations of tissue plasminogen activator as well as

of PAI-1,20 which occurs within hours of the administration of the toxin. Expression of several of the proteins of coagulation and fibrinolysis, including thrombomodulin,18 tissue factor, 18 and PAI-1 is regulated by proinflammatory cytokines,24 and there is evidence that gene transcription of the coagulation inhibitors may be controlled by similar transcriptional regulators to those that regulate cytokines.23 NF␬B recognition sequences can compete for the binding sites of a range of inflammation-induced transcriptional regulators that attach to the PAI-1 gene promoter.24 This linking of coagulation and inflammation suggests that the two pathways have evolved as interdependent mechanisms in the host response to infectious stimuli.27,28 Coagulation inhibitory pathways and those controlling fibrinolysis seem to be critical for down-regulating some of the adverse inflammatory effects initiated by sepsis.25,27 In animals, infusion of activated protein C is associated with improved survival of septic shock.29 Furthermore, inhibition of the tissue factor pathway by infusion of a specific inhibitor is also associated with improved outcome in experimental disseminated intravascular coagulation.30 Given the accumulation of data supporting an interdependence of the coagulation, fibrinolytic, and inflammatory pathways, our results (which establish a link between a functional polymorphism in the PAI-1 gene and outcome of meningococcal sepsis) may be less surprising than at first sight. Although disseminated intravascular coagulation is seen in virtually all forms of sepsis, the severity of the coagulopathy is especially prominent in meningococcal sepsis. Our finding, that a genetic predisposition to produce higher concentrations of PAI-1 in plasma in response to sepsis is associated with a poorer outcome of the disease, strongly suggests that over-production of this inhibitor of fibrinolysis may be an important event leading to the widespread intravascular thrombosis found in patients with septic shock and purpura. If the pathophysiology of fulminant purpura in meningococcal disease is indeed associated with a defective fibrinolytic response, owing to the inactivation of tissue plasminogen activator by PAI-1, measures to reduce the concentration of PAI-1 or to increase that of tissue plasminogen activator may be beneficial in treating the disease. PAI-1 is also a potent inhibitor of the protein C pathway.22 The association of deficiency in protein C or protein S with sepsis with fulminant purpura suggests that the protein C pathway may be important in preventing the dermal vascular thrombosis typical of this disease.14 Defective function of both fibrinolysis and the protein C anticoagulant pathway may be consequences of the overproduction of PAI-1. Further studies will be required to establish how high concentrations of PAI-1 are involved in the severity of meningococcal disease. Contributors P Hermans contributed to the design, writing, and analysis of the work, and undertook the genotyping. M Hibberd contributed to the organisation of the study, analysis of the data, and writing of the paper. R Booy helped to establish the meningococcal genetic study, to analyse the data, and to write the paper. O Daramola undertook the laboratory work and contributed to the analysis of the data. J Hazelzet contributed to the design of the study, analysis of clinical information, and writing of the paper. R de Groot initiated the study, and contributed to the analysis and writing of the paper. M Levin initiated the genetic study, oversaw the project, and contributed to the analysis and writing of the paper. The Meningococcal Research Group was responsible for collection and analysis of the clinical and laboratory information and samples from patients and controls.

Meningococcal Research Group The following members of the group participated in this study: E D de Kleijn, R F Kornelisse, M Sluijter, M H Suur, S Nadel, R Galassini, P Habibi, and J Britto.

Acknowledgments We thank A Panahloo (Department of Medicine, University College London Medical School at Whittington Hospital, London, UK) for providing control DNA samples with known PAI-1 genotypes. This work was supported by the Meningococcal Research Foundation; R Booy was a Wellcome Trust training fellow.

References 1

2

3

4

5

6

7

8

9

10

11 12

13

14

Achtman M. Global epidemiology of meningococcal disease. In: Cartwright K, ed. Meningococcal disease. Chichester: John Wiley & Sons, 1995: 159–77. Nadel S, Levin M, Habibi P. Treatment of meningococcal disease in childhood. In: Cartwright K, ed. Meningococcal disease. Chichester: John Wiley & Sons, 1995: 207–43. Brandtzaeg P, Kierulf P, Gaustad P, et al. Plasma endotoxin as a predictor of multiple organ failure and death in systemic meningococcal disease. J Infect Dis 1989; 159: 195–204. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T. The complex pattern of cytokines in serum from patients with meningococcal septic shock: association between interleukin 6, interleukin 1, and fatal outcome. J Exp Med 1989; 169: 333–38. Speer CP, Rethwilm M, Gahr M. Elastase-alpha 1-proteinase inhibitor: an early indicator of septicemia and bacterial meningitis in children. J Pediatr 1987; 111: 667–71. Osterud B, Flaegstad T. Increased tissue thromboplastin activity in monocytes of patients with meningococcal infection: related to an unfavourable prognosis. Thromb Haemost 1983; 49: 5–7. Powars DR, Rogers ZR, Patch MJ, McGehee WG, Francis RB. Purpura fulminans in meningococcemia: association with acquired deficiencies of proteins C and S. N Engl J Med 1987; 317: 571–72. Brandtzaeg P, Mollnes TE, Kierulf P. Complement activation and endotoxin levels in systemic meningococcal disease. J Infect Dis 1989; 160: 58–65. Hazelzet JA, Risseeuw-Appel IM, Kornelisse RF, et al. Age-related differences in outcome and severity of DIC in children with septic shock and purpura. Thromb Haemost 1996; 76: 932–38. Brandtzaeg P, Sandset PM, Joo GB, Ovstebo R, Abildgaard U, Kierulf P. The quantitative association of plasma endotoxin, antithombin, protein C, extrinsic pathway inhibitor and fibrinopeptide A in systemic meningococcal disease. Thromb Res 1989; 55: 459–70. Duncan A. New therapies for severe meningococcal disease but better outcomes? Lancet 1997; 350: 1565–66. Kuppermann N, Inkelis SH, Saladino R. The role of heparin in the prevention of extremity and digit necrosis in meningococcal purpura fulminans. Pediatr Infect Dis J 1994; 13: 867–73. Heyderman RS. Klein NJ, Shennan GI, Levin M. Deficiency of prostacyclin production in meningococcal shock. Arch Dis Child 1991; 66: 1296–99. Smith OP, White B, Vaughan D, et al. Use of protein-C concentrate, heparin, and haemofiltration in meningococcus-induced purpura fulminans. Lancet 1997; 350: 1590–93.

15 Fourrier F, Lestavel P, Chopin C, et al. Meningococcemia and purpura fulminans in adults: acute deficiencies of proteins C and S and early treatment with antithrombin III concentrates. Intensive Care Med 1990; 16: 121–24. 16 Zenz W, Muntean W, Gallistl S, Zobel G, Grubbauer HM. Recombinant tissue plasminogen activator treatment in two infants with fulminant meningococcemia. Pediatrics 1995; 96: 144–48. 17 Drapkin MS, Wisch JS, Gelfand JA, Cannon JG, Dinarello CA. Plasmapheresis for fulminant meningococcemia. Pediatr Infect Dis J 1989; 9: 399–400. 18 Moore KL, Andreoli SP, Esmon NL, Esmon CT, Bang NU. Endotoxin enhances tissue factor and suppresses thrombomodulin expression of human vascular endothelium in vitro. J Clin Invest 1987; 79: 124–30. 19 Engebretsen LF, Kierulf OP, Brandtzaeg P. Extreme plasminogen activator inhibitor and endotoxin values in patients with meningococcal disease. Thromb Res 1986; 42: 713–16. 20 Suffredini AF, Harpel PC, Parrillo JE. Promotion and subsequent inhibition of plasminogen activation after administration of intravenous endotoxin to normal subjects. N Engl J Med 1989; 320: 1165–72. 21 Pannekoek H, Veerman H, Lambers H, et al. Endothelial plasminogen activator inhibitor-1: a new member of the serpin gene family. EMBO J 1986; 5: 2539–44. 22 Gladson CL, Schleef RR, Binder BR, Loskutoff DJ, Griffin JH. A comparison between activated protein C and des-1-41-light chainactivated protein C in reactions with type 1 plasminogen activator inhibitor. Blood 1989; 74: 173–81. 23 Ryan MP, Kutz SM, Higgins PJ. Complex regulation of plasminogen activator inhibitor type-1 (PAI-1) gene expression by serum and substrate adhesion. Biochem J 1996; 314:1041–46. 24 Dawson SJ, Wiman B, Hamsten A, Green F, Humphries S, Henney AM. The two allele sequences of a common polymorphism in the promoter of the plasminogen activator inhibitor-1 (PAI-1) gene respond differently to interleukin-1 in HepG2 cells. J Biol Chem 1993; 268: 10739–45. 25 Kornelisse RF, Hazelzet JA, Savelkoul HFJ, et al. The relationship between plasminogen activator inhibitor-1 and pro-inflammatory and counter-inflammatory mediators in children with meningococcal septic shock. J Infect Dis 1996; 173: 1148–56. 26 Derkx B, Wittes J, McCloskey R, and the European Pediatric Meningococcal Septic Shock Trial Study Group. Randomized, placebo-controlled trial of HA-1A, a human monoclonal antibody to endotoxin, in children with meningococcal septic shock. Clin Infect Dis 1999; 28: 770–77. 27 Esmon CT, Taylor FB, Snow RT. Inflammation and coagulation: linked processes potentially regulated through a common pathway mediated by protein C. Throm Haemost 1991; 66: 160–65. 28 Muta T, Iwanaga S. The role of hemolymph coagulation in innate immunity. Curr Opin Immunol 1996; 8: 41–47. 29 Taylor FB Jr, Chang A, Esmon CT, D’Angelo A, Vigano-D’Angelo S, Blick KE. Protein C prevents the coagulopathic and lethal effects of Escherichia coli infusion in the baboon. J Clin Invest 1987; 79: 918–25. 30 Day KC, Hoffman LC, Palmier T, et al. Recombinant lipoproteinassociated coagulation inhibitor inhibits tissue thromboplastin-induced intravascular coagulation in the rabbit. Blood 1990; 76: 1538–45.