Behavior of protein S during long-term oral anticoagulant therapy

It has recently been reported that a natural anticoagulant, protein S (PS), is depressed during oral anticoagulation. Since more detailed information is required from the clinical standpoint, we measured plasma levels of PS [both total and free (not complexed) PS antigen], C4b-binding protein (C4bp) and other vitamin K-dependent proteins (factors II, VII, IX, X and protein C) in 60 plasma samples from patients on long-term oral anticoagulant therapy with warfarin. Together with the reduction of other vitamin K-dependent plasma proteins, PS decreased during warfarin treatment, being dependent on the intensity of the therapy. A considerable variation in plasma PS levels was also observed among individuals with a similar intensity of anticoagulation. Plasma concentration of C4bp was closely correlated with total PS level, and free PS/total PS ratio was independent of thrombotest values. These findings indicate that long-term oral anticoagulant therapy results in the suppression of the synthesis of PS, and that its reduction is on the whole balanced with C4bp and vitamin K-dependent coagulation factors. It was suggested that the metabolism of C4bp might be regulated by the plasma PS level, although this hypothesis needs further exploration.     

Protein C, an anticoagulant protein, is increased in healthy volunteers and surgical patients after treatment with stanozolol

On both oral and intramuscular administration, the anabolic steroid stanozolol was found to increase protein C antigen concentrations in circulating blood. In fourteen healthy young volunteers (who received stanozolol orally, dose 10 mg/day) the average increase was 1.5-1.6 times the normal concentrations after 3-6 weeks' treatment and was accompanied by more moderate increases in the other vitamin K-dependent factors II, IX and X to 1.4, 1.4 and 1.2 times their normal concentration respectively. However, there was no change in factor VII. In sixteen elderly surgical patients, intramuscular injection (50 mg) one day prior to surgery induced a moderate increase within 24 hours (to 1.11 times the pretreatment concentration) and seven days after operation (to 1.19 times), and reduced the postoperative fall in protein C. Stanozolol administration seems to be a promising pharmacological method for increasing anticoagulant protein C levels in congenital and acquired deficiencies.     

The relative importance of the factors II, VII, IX and X for the prothrombinase activity in plasma of orally anticoagulated patients

The individual importance of each of the four vitamin K-dependent clotting factors on the generation of prothrombinase activity in the plasma of orally anticoagulated patients has been investigated. Addition of purified factors VII, IX or X to plasma from deeply anticoagulated patients (International Normalized Ratio 2.8-4.8) did not influence the amount of prothrombinase activity or the amount of thrombin formed. Only the prothrombin level in the plasma determines the course of thrombin generation. Addition of increasing amounts of purified factor II, VII, IX or X to plasmas deficient in respectively factor II, VII, IX or X showed that the prothrombinase activity increases linearly with the concentration of factor II added and that the concentration below which the factors VII, IX and X start to have a measurable effect on prothrombinase activity are 5%, 20%, and 30%, respectively. Half maximal prothrombinase activity was found at about 1% factor VII, 5% factor IX and 8% factor X respectively. From these observations we conclude that primarily the variation in factor II level determines thrombin generation and hence presumably the antithrombotic effect of oral anticoagulant therapy. It therefore seems likely that, for the control of oral anticoagulant therapy, tests that reflect factor II activity would be suitable.     

A new method for the estimation of protein C by ELISA

A new solid phase enzyme-linked immunosorbent assay (ELISA) was developed for the quantitation of human protein C antigen. Anti-protein C F(ab′)2 fragments were adsorbed to polystyrene plates. The binding of serial dilutions of control or test plasma, containing protein C, was detected by incubation with peroxidase-labeled anti-protein C-IgG followed by the addition of hydrogen peroxyde and 0-phenylenediamine. This ELISA is specific, sensitive (detection limit : 0.02 %) and accurate (variation coefficient : 3 to 10 %). When results are compared to those obtained by the Laurell technique (electroimmunodiffusion, EID), the correlation coefficient is 0.95 in all tested plasmas.

Protein C antigen was measured by ELISA and EID in plasma from 40 controls, 14 patients with congenital protein C deficiency, 15 patients with liver cirrhosis and 40 dicoumarol-treated cases. In normal plasma, protein C ranged from 70 to 126 %. In congenital deficiency, protein C was between 35 and 58 % in 13 cases and 9 % in one of them. In patients with liver cirrhosis and dicoumarol-treated cases, levels of protein C antigen were compared to those of other vitamin K dependent factors, i.e. Factors II and IX measured by EID and Factor X assayed by EID and ELISA. In liver cirrhosis, the amount of protein C was significantly lower than that of Factors II, IX and X. In short-term and long-term dicouma-rol-treated patients, the highest correlation (r = 0.72) was observed between protein C and Factor X levels. In the plasma of patients undergoing oral anticoagulant therapy, protein C decreased more rapidly than Factors X or II and migrated in presence of calcium as a double peak, one with a normal mobility and one more anodal corresponding to the non carboxylated form of protein C.     

Anti- and procoagulant activities in factor VII-deficient subjects

The clinical feature in patients with congenital factor VII deficiency is in part dependent on the underlying genetic defect, but the mechanisms influencing the genotype–phenotype correlation remain to be fully elucidated. In addition, thromboembolic events have been reported. Compensatory mechanisms involving vitamin K-dependent factors have been suggested. We have measured anticoagulant activities in 25 factor VII-deficient subjects (factor VII activity ≤36%) and 23 age-matched controls and correlated these to the vitamin K-dependent procoagulant activities. Two of the patients had a history of thromboembolism. The factor VII-deficient patients were found to have a significantly lower protein C activity than the controls [0.84 U/ml (95% CI 0.78; 0.89) vs. 0.98 U/ml (95% CI 0.91; 1.05), P=.004]. In addition, the protein C activity was correlated to that of factor VII (r=.36; P=.014), factor IX (r=.45; P=.002) and factor X (r=.50; P=.0006), respectively. The level of prothrombin fragment 1+2 was correlated to the protein C (r=.40; P=.012) and to the factor VII activity (r=.42; P=.011). No differences between patients and controls were seen regarding total and free protein S, antithrombin, plasminogen activator inhibitor-1 (PAI-1) and tissue factor pathway inhibitor (TFPI). Seven of the patients were found to have the Factor V Leiden mutation, but none of them had experienced any thromboembolic event. The present data support the notion that compensatory hemostatic mechanisms might exist in that the protein C activity was found to be decreased in the factor VII-deficient subjects. Whether this could influence the clinical feature, including the risk of thromboembolic events in association with replacement therapy, remains to be evaluated.     

Protein C and protein S levels in uremic patients before and after dialysis

The inhibitory capacity of the natural protein C (PC)/protein S (PS) system was evaluated measuring both the functional activity and the antigen level of both these inhibitors in 30 uremic patients before and after a dialytic treatment and in 30 healthy normal volunteers. PC functional activity was determined by two methods, one clotting and one chromogenic. PS antigen level was measured both as free protein and as total content. Unlike previous authors, we found that PC functional activity and the antigen level were normal in patients before dialysis, with a significant increase after. PS functional activity and free and total antigen levels were all normal before dialysis, and all except free antigen showed a significant post-treatment rise.     

Increased activity of vitamin K-dependent clotting factors in human hyperlipoproteinaemia - association with cholesterol and triglyceride levels

Levels of blood coagulation factors, cholesterol and triglyceride were measured in human plasma. Prothrombin was significantly elevated in type IIa hyperlipidaemia; prothrombin and factors VII, IX and X in type IIb; and prothrombin and factors VII and IX in type V. Multiple regression analysis showed significant correlation between the levels of these plasma lipids and the vitamin K-dependent clotting factors (prothrombin, factors VII, IX and X). Higher cholesterol levels were associated with higher levels of prothrombin and factor X while higher triglyceride levels were associated with higher levels of these as well as factors VII and IX. Prothrombin showed a significant cholesterol-triglyceride interaction in that higher cholesterol levels were associated with higher prothrombin levels at all levels of triglyceride, with the most marked effects in subjects with higher triglyceride levels. Higher prothrombin levels were noted in subjects with high or moderately elevated (but not low) cholesterol levels. Ultracentrifugation of plasma in a density of 1.21 showed activity for prothrombin and factors VII and X only in the lipoprotein-free subnatant fraction. Thus, a true increase in clotting factor protein was probably present. The significance of the correlation between levels of vitamin K-dependent clotting factors and plasma lipids remains to be determined.     

Correlation between different intensities of anti-vitamin K treatment and coagulation parameters

In order to study the effect of different intensities of anti-vitamin K treatment on coagulation parameters, 23 patients with venous thromboembolism were given, after the initial treatment period, warfarin at doses giving an International Normalised Ratio of 1.3–2.0 for 4 weeks, and of 1.1–1.3 for another 4 weeks. Blood samples were taken at the end of each of these periods and 4 weeks after the end of warfarin treatment. The vitamin K-dependent coagulation factors VII, IX, and X, as well as the inhibitor protein C and its cofactor protein S, all showed a highly significant correlation with treatment intensity. This was to some extent also true for the coagulation activation markers, prothrombin fragment 1+2 and thrombin–anti-thrombin complex. Ratios of pro- and anticoagulant factors in some instances showed a decrease at therapeutical (International Normalised Ratio) levels, and also sometimes with reduced warfarin treatment intensity. Taken together, our results encourage further research addressing issues of varying treatment intensity with warfarin and alternative methods for monitoring of anti-vitamin K treatment.     

Expression of protein S in the murine heart and cultured mouse cardiomyocytes is down-regulated by cytokines

Protein S (PS), a co-factor of activated protein C, is a vitamin K-dependent anticoagulant protein and is known to be produced extrahepatically. In the present study, the concentration of PS mRNA was determined tissue by tissue in the mouse, and it was high in lung, adrenal and heart as well as in liver. We further investigated the effects of lipopolysaccharide (LPS), tumor necrosis factor-alpha (TNF-alpha), and interleukin-1 (IL-1) on the PS mRNA expression in murine tissues in vivo. Although LPS and TNF-alpha significantly decreased the expression level of PS mRNA in all tissues examined (e. g., lung, liver, heart, and kidney) and the PS antigen level in plasma, the suppressive effect of IL-1 on PS gene expression was limited to heart. More specifically, considerable amounts of PS mRNA and antigen were expressed in a cultured mouse cardiomyocyte cell line, and again, treatment with IL-1 decreased the PS expression in these cells. These observations raise a possibility that the expression of cardiac PS may contribute to the regional anticoagulant potential in heart, and suggest that the decreased PS expression by cytokines may result in an increase in the systemic and/or regional prothrombotic potential under inflammatory conditions.     

Genetic regulation of plasma levels of vitamin K-dependent proteins involved in hematostatis: results from the GAIT Project. Genetic Analysis of Idiopathic Thrombophilia

Vitamin K-dependent proteins play a critical role in hemostasis. We have analysed the genetic and environmental correlations between measures of several vitamin K-dependent proteins in 21 Spanish extended families, including 397 individuals. Plasma functional levels of factors II, VII, IX, X, protein C and functional protein S were assayed in an automated coagulometer. Antigenic levels of total and free protein S were measured using an ELISA method. A maximum likelihood-based covariance decomposition analysis was used to assess the heritability of each trait and the genetic and environmental correlations between all possible pairs. All of the plasma levels had a significant genetic component (heritability) ranging from 22% to 52% of the phenotypic variance. Among the 28 possible pairs of genetic correlations, 18 were significant at a level of p < 0.05 and six exhibited a p-value between 0.05 and 0.10. Positive environmental correlation was observed for 25 of the pairs (p < 0.05). We conclude that genetic effects account for a large proportion of the observed phenotypic variation in vitamin K-dependent proteins. Some of the genes appear to pleiotropically influence all of these traits, since most pairs of phenotypes exhibit significant genetic correlation. However, since these phenotypes show a high degree of environmental correlation, it is also likely that the same environmental factors influence them co-jointly.     

Purification of rabbit, rat and mouse protein C with the use of monoclonal antibody to human protein C, PC01

Ca(++)-dependent monoclonal antibody specific to gamma-carboxyglutamic acid (Gla) domain of protein C was produced. It did not cross-react to the other vitamin K-dependent plasma proteins but to protein C of the other species. Using this monoclonal antibody, PC01, rabbit (170 micrograms), rat (60 micrograms) and mouse (40 micrograms) protein Cs were isolated from 100 ml of their plasma by affinity chromatography. All of these protein Cs were two chain form linked by disulfide bond as well as human protein C and activated by thrombin-thrombomodulin complex. Rat and mouse protein Cs showed similar characters to human protein C. On the other hand rabbit protein C had different M(r) of heavy and light chains and showed lower anticoagulant activity compared with human protein C.     

Neonatal protein C: molecular composition and distribution in normal term infants

Protein C (PC) is a vitamin K-dependent serine protease which functions as the central regulatory protein with both anticoagulant and profibrinolytic properties. The PC levels in healthy term newborns are approximately one third of adult levels. Severely decreased levels of PC are seen in sick term and preterm infants. These neonates appear to have an increased incidence of thrombosis. Undetectable levels of PC are found in homozygous PC deficient infants with DIC and purpura fulminans symptoms. In this present study we report the composition and distribution of PC in term newborn and compare the results with adult values. Plasma was obtained from placental cord blood of 20 healthy term (38-42 weeks gestation) infants. PC was immunopurified, run on SDS-PAGE, and immuno-blotted. The composition of the PC molecule in neonatal plasma is identical to that seen in adults. Using densitometry to determine the distribution of the PC components, we observed a 2-fold increase in single chain PC in the neonate as compared to the adult. In the neonate, there was an inverse correlation between the level of total PC antigen and the amount of single chain. These findings suggest the possibility that the processing of PC may be developmentally influenced.     

Effect of recombinant human FVIIA on warfarin-induced bleeding in rats

Recombinant human factor VIIa (rFVIIa) was given to warfarin-treated rats. One and two warfarin treatments reduced endogenous factor X levels to about 10% and <5%, respectively. The reduction of plasma levels of vitamin K-dependent coagulation factors was accompanied by a treatment-dependent prolongation of activated partial thromboplastin time (APTT) and prothrombin time (PT) as well as by increased bleeding time and blood loss in the rat tail bleeding test. rFVIIa 50 μg/kg and 250 μg/kg normalized PT and shortened APTT in rats given warfarin once. Bleeding was completely normalized by rFVIIa 250 μg/kg. In rats given warfarin twice rFVIIa 250 μg/kg shortened PT but had no effect on APTT, whereas the effect on bleeding was variable. The elimination half life of rFVIIa in rats was found to be 30–40 minutes. The study indicates that rFVIIa may be useful in patients with bleeding due to decreased levels of vitamin K-dependent coagulation factors.     

The anticoagulant properties of a modified form of protein S

Protein S (PS) is a vitamin K-dependent anticoagulant that acts as a cofactor to activated protein C (APC). To date PS has not been shown to possess anticoagulant activity in the absence of APC. In this study, we have developed monoclonal antibody to protein S and used to purify the protein to homogeneity from plasma. Affinity purified protein S (PSM), although identical to the conventionally purified protein as judged by SDS-PAGE, had significant anticoagulant activity in the absence of APC when measured in a factor Xa recalcification time. Using SDS-PAGE we have demonstrated that prothrombin cleavage by factor Xa was inhibited in the presence of PSM. Kinetic analysis of the reaction revealed that PSM competitively inhibited factor Xa mediated cleavage of prothrombin. PS preincubated with the monoclonal antibody, acquired similar anticoagulant properties. These results suggest that the interaction of the monoclonal antibody with PS results in an alteration in the protein exposing sites that mediate the observed anticoagulant effect. Support that the protein was altered was derived from the observation that PSM was eight fold more sensitive to cleavage by thrombin and human neutrophil elastase than conventionally purified protein S. These observations suggest that PS can be modified in vitro to a protein with APC-independent anticoagulant activity and raise the possibility that a similar alteration could occur in vivo through the binding protein S to a cellular or plasma protein.     

Circadian variations in natural coagulation inhibitors protein C, protein S and antithrombin in healthy men: a possible association with interleukin-6

Recent observations describe an increase in platelet aggregability and a decrease in fibrinolytic activity in the early morning hours. To determine whether anticoagulant proteins also show such a circadian variation we measured protein C (PC), protein S (PS), antithrombin (AT) and heparin cofactor-II (HC-II) levels on venous plasma samples taken from 10 healthy men at three-hour intervals throughout a 24-hour period. To investigate the possible temporal mapping of circadian periodicity, we also measured plasma levels of beta-thromboglobulin (beta-TG) as an indicator of platelet activation, and interleukin-6 (IL-6) as one of the possible regulatory factors that drive this rhythm. Blood samples were drawn at 6 a.m., 9 a.m., noon, 3 p.m., 6 p.m., 9 p.m. and midnight. PC, IL-6 and beta-TG were measured by ELISA, PS and AT by latex immune assay and HC-II by chromogenic substrate method. A significant circadian variation was found in PC, PS, AT, beta-TG and IL-6, but not in HC-II levels. PC, PS, IL-6 and beta-TG were at their peaks at 6 a.m., and nadirs at a time from noon to midnight. AT peak was at 6 p.m. and nadir at noon. The regression of PS on IL-6 was significant. Although the fluctuations of PS and AT were within the normal ranges during the day, some PC levels of two subjects were below the lower normal limit (0.70). These data indicate that PC, PS, and AT show a marked circadian periodicity as the other components of the blood coagulation and fibrinolytic system do. The similar trends in plasma concentrations of PC, PS, beta-TG and IL-6 may be coincidental, but could reflect a common regulatory mechanism or an effect on each other. The clinical implications of these physiological changes in coagulation inhibitors and the role of IL-6 in the anticoagulant response deserve further studies.     

Hereditary protein S deficiency and venous thrombo-embolism. A study in three Dutch families

Protein S, a vitamin K-dependent coagulation factor, is involved in the regulation of the anticoagulant activity of activated protein C. Using an immunoradiometric assay for total protein S in plasma we identified 14 patients (7 male and 7 female) in three unrelated Dutch families as fulfilling the criteria for an isolated protein S deficiency. In 9 patients who were not receiving oral anticoagulant treatment the mean total protein S antigen concentration was 0.50 +/- 0.08 U/ml (+/- S.D.) and the calculated free protein S concentration was 0.15 +/- 0.01 U/ml (+/- S.D.). In the five patients who were on oral anticoagulant treatment the mean total protein S antigen was 0.23 +/- 0.05 U/ml (+/- S.D.). Seven of the 14 patients had a history of venous thromboembolism occurring at a mean age of 25 years and often without an apparent cause. Protein S deficiency is inherited as an autosomal dominant trait.     

Alpha 2-macroglobulin enhances prothrombin activation and thrombin potential by inhibiting the anticoagulant protein C/protein S system in cord and adult plasma

Protein S (PS) is a vitamin K-dependent plasma protein and serves as a cofactor for the anticoagulant activities of activated protein C (APC). We investigated the effects of different PS concentrations on prothrombin activation and thrombin generation in cord and adult plasma containing APC and different amounts of alpha 2-macroglobulin (a2-M). Prothrombin activation was assessed by monitoring the time-course of prothrombin fragment 1+2 (F1+2) generation. Thrombin generation curves were determined by means of a subsampling technique using the chromogenic substrate S-2238. We demonstrate a dose-dependent inhibition of the anticoagulant action of PS by a2-M: suppression of F1+2 and thrombin generation due to addition of PS was stronger in plasma containing low amounts of a2-M than in plasma with elevated a2-M levels. Since no complex formation between a2-M and PS was observed by means of SDS-PAGE, we attribute decreased anticoagulant action of PS at high a2-M levels to enhanced complex formation between APC and a2-M. Thereby, APC is subtracted from its cofactor PS, resulting in suppressed formation of the anticoagulant APC/PS complex. Thus, our data suggest that a2-M, besides its well-known anticoagulant effects, also acts as a procoagulant by suppressing the formation of the anticoagulant APC/PS complex. Our findings have implications particularly on thrombin generation and inhibition in cord plasma, since a2-M levels in newborns are elevated over adult values and the antithrombotic APC/PS pathway is up-regulated at birth. Therefore, elevated levels of a2-M might restrict the up-regulation of the APC/PS pathway.     

Immunoradiometric assay for the calcium-stabilized conformation of human protein S

Rabbit polyclonal anti-protein S serum was fractionated with immobilized human protein S to establish solid-phase immunoradiometric assays recognizing Ca(II)-dependent and NonCa(II)-dependent epitopes of human protein S. The two assays were specific for PS:Ca(II)Ag and PS:NonCa(II)Ag and highly sensitive with a lower limit of detection of about 2.5 ng/ml. PS:Ca(II)Ag and PS:NonCa(II)Ag levels were measured in immunopurified protein S, thrombin-modified protein S and chymotrypsin-cleaved protein S. Only in chymotrypsin-cleaved protein S an important discrepancy between the two antigen levels was observed. Ranges for the concentration of PS:Ca(II)Ag and PS:Non-Ca(II)Ag and their ratio were established in plasma of healthy individuals (0.92 +/- 0.13 U/ml, 0.98 +/- 0.21 U/ml, 0.96 +/- 0.17, respectively). In a group of patients using oral anticoagulant therapy the ratio PS:Ca(II)Ag/PS:NonCa(II)Ag decreased at increasing intensity of anticoagulation suggesting the presence of sub- and noncarboxylated protein S molecules. In plasma of patients with a hereditary type I protein S deficiency PS:Ca(II)Ag and PS:NonCa(II)Ag were reduced to the same extent: mean ratio 1.02 +/- 0.12 in the group not on oral anticoagulant treatment and 0.94 +/- 0.10 in the group on oral anticoagulant therapy. Analysis of patients with a history of unexplained thrombo-embolic disease did not reveal individual patients with a PS:Ca(II)Ag/PS:NonCa(II)Ag ration below the lower limit of the normal range (mean ratio 1.05 +/- 0.17), suggesting that the frequency of genetic protein S variants with defects in the Ca(II)-stabilized conformation is very low.     

Normal reference ranges of antithrombin,protein C and protein S: Effect of sex,age and hormonal status

Introduction

Antithrombin (AT), protein C (PC) and protein S (PS) deficiencies are risk factors for venous thromboembolism. Overlapping values between heterozygous carriers and normal individuals often make a correct classification of a deficiency difficult. The aim of this study was to investigate the effect of sex, age, menopause and hormone therapy on natural anticoagulant plasma levels in a large group of healthy individuals, and to evaluate the need of separate reference ranges.

Materials and Methods

AT and PC were measured with a chromogenic assay, antigenic free PS with an ELISA test. To evaluate the effect of sex, age, oral contraception, hormonal status (and their interaction) on AT, PC and PS levels, linear regression models were used. Biological relevance and the value of the normal deviate z were chosen as rules to decide for separate reference ranges.

Results

The study population consisted of 1837 healthy adult individuals (741 men, 1096 women), aged 18-85 years (median age: 44 years). In men AT levels decreased after the age of 50 years. Men had higher levels of PS than women, particularly at young ages. In women, after correction for menopause, only PC levels increased with age. Menopause affected AT and PS, but not PC levels. Oral contraceptive intake was associated with a decrease of AT and PS, and an increase of PC levels.

Conclusions

For AT, PC and PS, sex- and age-specific normal reference ranges can be useful, in order to better discriminate true carriers of a natural anticoagulant deficiency.