- Email: [email protected]
Nutritional vitamin K-intake and urinary γ-carboxyglutamate excretion in the rat
Biochimica et Biophysica Acta 1334 Ž1997. 44–50
Nutritional vitamin K-intake and urinary g-carboxyglutamate excretion in the rat Alexandra M. Craciun, Monique M.C.L. Groenen-van Dooren, Cees Vermeer
)
CardioÕascular Research Institute Maastricht and Department of Biochemistry, UniÕersity of Limburg, P.O. Box 616, 6200 MD Maastricht, The Netherlands Received 11 March 1996; revised 9 July 1996; accepted 17 July 1996
Abstract Using the rat as an experimental animal model we have found that prothrombin synthesis reaches its maximal level at a relatively low dietary vitamin K intake. At still higher vitamin K intakes, however, the urinary Gla-excretion was substantially increased, showing a different vitamin K requirement for liver and extrahepatic tissues. The increased urinary Gla-excretion was found for both phylloquinone and menaquinone-4, but not for menaquinone-8, which questions the bioavailability of higher menaquinones for extrahepatic tissues. A discrepancy was found between effects of nutritional vitamin K-deficiency and treatment with a vitamin K-antagonist Žbrodifacoum.. With both regimens plasma prothrombin rapidly decreased to well below 10% of the starting values, but in case of K-deficiency urinary Gla had hardly decreased in 7 days, whereas after 3 days of brodifacoum treatment Gla-excretion had decreased to 17% of the starting values. An explanation for this observation is that prothrombin procoagulant activity does not decrease proportional to the prothrombin Gla-content, but that a wide range of undercarboxylated prothrombins have lost nearly all activity. During vitamin K-deficiency the remaining low levels of vitamin K would mainly give rise to undercarboxylated prothrombin, whereas during brodifacoum treatment only non-carboxylated prothrombin is formed. It seems plausible that in the latter case the urinary Gla originates from proteins with long half-life times, such as the bone Gla-proteins. Keywords: Vitamin K; Phylloquinone; Menaquinone; g-Carboxyglutamate; Coagulation; Posttranslational
1. Introduction Vitamin K is a group name for a number of related compounds which all have a naphthoquinone ring structure, but which differ from each other by the length and degree of saturation of their aliphatic side
Abbreviations: Gla, g-carboxyglutamate Corresponding author. Fax: q31 43 3670992.
)
chain at the 3-position w1x. The three K-vitamers which were studied in this paper are: phylloquinone, menaquinone-4, and menaquinone-8. Phylloquinone and menaquinone-4 are different with respect to their degree of saturation, whereas menaquinone-4 and menaquinone-8 differ in their aliphatic side chain length. The only known function of vitamin K in mammals is to act as a coenzyme for the endoplasmic enzyme g-glutamylcarboxylase during the posttranslational conversion of protein-bound glutamate
0304-4165r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 0 4 - 4 1 6 5 Ž 9 6 . 0 0 0 7 3 - 6
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
residues into g-carboxyglutamate ŽGla. w2,3x. The daily requirement of vitamin K is extremely low, because a second enzyme ŽKO-reductase. is capable of recycling the vitamin so that one molecule may catalyze the formation of several thousand Gla-residues w2,4x. A number of coumarin derivatives Že.g., warfarin and brodifacoum. inhibit the recycling of vitamin K by blocking the KO-reductase, thus rapidly exhausting the pool of bioavailable vitamin K w5,6x. Once it has been formed Gla cannot be metabolized or re-used, so eventually all Gla is excreted stoichiometrically via the urine, most of it as a free amino acid and less than 5% in a peptide-bound form w7–9x. Hence urinary Gla levels may reflect the turnover of the total body pool of Gla proteins; in a steady state situation the urinary Gla excretion will even be a measure for the total Gla production. By comparing plasma prothrombin and urinary Gla under different conditions an impression may be obtained of the extent and vitamin K-requirement of extra-hepatic carboxylation events. The enzymes of the vitamin K-cycle Ž i.e., carboxylase and reductase. have been discovered in the liver, but afterwards they were detected in most tissues and cell types w10–13x. The hepatic system is involved in the production of several blood coagulation factors, the most abundant of which is prothrombin. Osteoblasts in bone produce osteocalcin and matrix Gla-protein, two proteins which are structurally not related to the hepatic Gla-proteins w14x. Recent studies in humans have indicated that the nutritional vitamin K-requirement for complete carboxylation of osteocalcin may be higher than that for the carboxylation of prothrombin and related coagulation factors w15,16x. In this paper we have made use of the rat as an experimental model, and we have assessed the minimal dietary vitamin K intake at which the production of the hepatic Gla-protein prothrombin takes place at its maximal rate. To exclude small variations due to interindividual differences all animals were adapted for 3 weeks to a diet containing 3 times the minimal vitamin K requirement. In this model we have investigated the effect on the plasma prothrombin concentration and the urinary Gla excretion of a very high intake of various K vitamers, that of a nutritional vitamin K-deficiency, and finally the effect of administration of the vitamin K-antagonist brodifacoum.
45
2. Materials and methods 2.1. Chemicals Phylloquinone Žvitamin K 1 ., D,L-g-carboxyglutamic acid ŽGla., and orthophthaldialdehyde ŽOPA. were purchased from Sigma ŽSt. Louis, MO.; menaquinone-4 and menaquinone-8 were kind gifts from Hoffmann-La Roche ŽBasel, Switzerland.. The non-ionic surfactant HCO-60 Žpolyoxyethylene hydrogenated castor oil derivatives. was purchased from Nikko Chemicals ŽTokyo, Japan.. Brodifacoum was a gift from Sorex ŽCheshire, UK. and it was dissolved to a final concentration of 0.25 mgrml in a buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.8 and 1% Žwrv. HCO-60 shortly before use. The various K-vitamers were dissolved in the same buffer system to a final concentration of 10 m molrml. All other chemicals were of analytical grade or better. 2.2. Animals and diets All studies were performed in male rats of the Lewis strain, which were 12 weeks old when entering in the experiment. The animals were housed in individual metabolic cages equipped for separate collection of urine and faeces ŽTecniplast 170102 cages, B and E Industry, Utrecht, The Netherlands., with a 12-h light-dark cycle and controlled temperature Ž 20 " 28C. and humidity Ž50 " 10%.. They had free access to food and water. The following foods were used: 1. A commercial rat diet Ž SRM-A, Hope Farms, Woerden, The Netherlands., the vitamin K content of which was found to be 5 mg of menadione and 2 mg of phylloquinone per kg of food. 2. A radiated Ž 0.9 Mrad. vitamin K-deficient diet ŽHope Farms, Woerden, The Netherlands., which was mixed with cooked and dried white rice in a 1:1 Žwrw. ratio prior to use w17x. This food was checked not to contain detectable amounts of vitamin K. 3. The same vitamin K-deficient diet, mixed with cooked and dried rice, to which 20 m g of menaquinone-4 per g of food were added. All food was powdered with a blender and mixed in a professional food processor before use. To ensure a controlled vitamin K intake the vitamers were administered orally, via a syringe equipped with a
46
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
plastic cannula, which was protruded into the esophagus. All rats were weighed regularly, and it was checked that their body weights increased by about 1 g per day, and that the various groups did not differ with respect to mean body weight gain. Blood Ž 0.5 ml. was collected in 0.05 ml of 0.14 M trisodium citrate after puncture of the tail vein. The blood samples were centrifuged for 15 min at 3 000 = g and the plasma was frozen at y808C until use. Urines were collected during the last 24 h of each measuring point, the volumes were recorded. Subsequently the samples were centrifuged for 10 min at 5 000 = g and the supernatants were frozen at y808C until use. 2.3. Urinary Gla detection The urine samples were diluted with distilled water between 4 and 160 times Žas required to stay within the normal detection range., and the diluted samples were analyzed according to the procedure of Kuwada and Katayama w18x, using a Separations high precision HPLC pump Žmodel 300., a Kratos Spectraflow 980 detector and a Nucleosil 100-5 SB anion exchange resin Ž5 m m particles, Macherey-Nagel, Duren, Germany. at a flow rate of 1 mlrmin. Refer¨ ence curves were prepared from samples with known concentrations of authentic Gla. The sensitivity of the assay was 0.3 m M. The intra- and interassay coefficients of variation were 2.2 and 6.4%, respectively.
and the Wilcoxon test for experiments with less animals per group. All data are expressed as mean values" S.E.
3. Results 3.1. Gla-excretion at the minimal Õitamin K intake required for normal blood coagulation Sixty rats were made vitamin K-deficient by feeding them with a commercial diet devoid of vitamin K for a period of 1 week. At that time blood was collected, and it was found that the circulating prothrombin concentration had decreased to 10–20% of the starting values. Subsequently the rats were subdivided into 10 groups, and varying Žknown. amounts of menaquinone-4 were added to the diet. This regime was continued for 2 weeks before blood was taken for another prothrombin determination. It was found that at a menaquinone-4 content of 6 m g Ž13.5 nmol. per g of food the prothrombin levels had reached their pre-treatment levels, and that they did not change at higher intakes Žsee Fig. 1.. Since the mean food consumption was 22 g per day, it may be calculated that the minimal dietary menaquinone-4 requirement was around 300 nmol per day. On the basis of these data it was decided that for all experiments described below the reference rats Ži.e., vitamin K-sufficient. would receive a diet containing 20 m g Ž45 nmol. of
2.4. Plasma prothrombin assay Prothrombin concentrations were determined in the one-stage assay using a coagulometer Ž KC-4, Amelung, Germany. , and a commercial thromboplastin preparation ŽThromborel S w . and clotting factor II deficient plasma Ž both from Behringwerke AG, Marburg, Germany.. Prothrombin concentrations were calculated with the aid of reference curve from pooled normal rat plasma. 2.5. Statistical analysis The software package SPSSrPCq version 5.0 ŽSPSS, Chicago, IL. was used for the statistical analysis and the 0.05 level of significance was chosen. The data were analyzed using the paired Student’s t-test for comparing groups of more than 20 animals,
Fig. 1. Menaquinone-4 requirement for normal prothrombin synthesis in the vitamin K-deficient rat. The food was administered ad libitum during a period of 2 weeks. Plasma prothrombin concentrations are expressed as a percentage of the starting values. Each point represents the mean value Ž"S.E.. for six rats.
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
47
Table 1 Effect of diet on plasma prothrombin and urinary Gla excretion Marker
Before treatment
After treatment
Difference
P-value
Prothrombin Ž%. Gla-excretion Žnmolrday. Glarprothrombin ratio
80.2 Ž"0.8. 1470 Ž"70. 18.4 Ž"1.
79.9 Ž"1.8. 1340 Ž"70. 17.0 Ž"1.
y0.28 Ž"1.7. y130 Ž"40. y1.4 Ž"0.5.
N.S. - 0.005 - 0.02
Measurements were made in rats before and after feeding for three weeks with the standardized control diet. Prothrombin was expressed relative to a reference pool of rat plasma, Gla-excretion was measured in 24-h urines. The significance of differences was calculated with the paired Student’s t-test. N.S. stands for not significant Ž P ) 0.05..
menaquinone-4 per g. This diet will be referred to as the ‘standardized control diet’. Reference rats were defined further in an experiment in which we have measured the plasma prothrombin concentration and the urinary Gla excretion in 25 animals before they entered the experiment and after they had been fed for 3 weeks with the standardized control diet. The data are given in Table 1, and demonstrate that plasma prothrombin had not changed by the switch from commercial food to the standardized control diet used for known vitamin K-intake. The urinary Gla-excretion had slightly decreased, however, which suggests that the Gla-excretion can be decreased without changing the plasma prothrombin concentration, and also that more vitamin K can be extracted from the commercial chow than from the standardized control diet. The divergence between urinary Gla excretion and plasma prothrombin concentration is visualized most clearly from the ratio between both variables. As can be seen from Table 1
this ratio had decreased significantly after the dietary change. 3.2. Effects of high doses of Õarious K-Õitamers on plasma prothrombin and on urinary Gla Thirty five rats were kept for 3 weeks on the standardized control diet, and subsequently they were subdivided into three groups of 10 and one of group of 5 animals. The first group Ž n s 10. continued with the standardized control diet as its sole source of nutritional vitamin K, the other three groups received a high oral dose Ž20 m moles. of vitamin K on each of three successive days. Three vitamers were studied: phylloquinone Ž n s 10., menaquinone-4 Ž n s 10., and menaquinone-8 Ž n s 5.. As is shown in Table 2, these high doses of vitamin K did not change the circulating prothrombin concentrations. On the other hand, urinary Gla excretion was substantially and significantly increased by both phylloquinone and
Table 2 Effect of high vitamin K intake on plasma prothrombin and urinary Gla Control n s 10
Excess K 1 n s 10
Excess MK-4 n s 10
Excess MK-8 ns5
Gla-excretion Žnmolrday. Before treatment After treatment Difference Statistical significance Ž P .
1464 Ž"178. 1451 Ž"165. y13 Ž"50. N.S.
1501 Ž"204. 1853 Ž"232. q352 Ž"110. 0.02
1372 Ž"280. 1875 Ž"406. q503 Ž"210. 0.01
1330 Ž"75. 1486 Ž"86. q156 Ž"70. N.S.
Prothrombin Ž%. Before treatment After treatment Difference Statistical significance Ž P .
77.2 Ž"1.8. 83.4 Ž"3.4. q6.2 Ž"7.5. N.S.
84.4 Ž"3.7. 89.9 Ž"5.1. q5.5 Ž"7.1. N.S.
85.9 Ž"4.5. 90.6 Ž"6.4. q4.7 Ž"8.3. N.S.
81.6 Ž"2.1. 89.2 Ž"3.8. q7.6 Ž"3.5. N.S.
K 1 stands for phylloquinone, MK for menaquinone. Statistical significance of differences was calculated with the Wilcoxon test. Further details are as in the legend to Table 1.
48
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
menaquinone-4, but not by menaquinone-8. Prolonged treatment with the high doses of phylloquinone and menaquinone-4 until day 7 did not change these data further Ždata not shown.. Unfortunately our supply was insufficient to continue the treatment with menaquinone-8. 3.3. Effects of dietary Õitamin K-deficiency and oral anticoagulants on plasma prothrombin and on urinary Gla Thirty rats were kept for 3 weeks on the standardized control diet, and subsequently they were subdivided into three groups of 10 animals each. In the first group, which served as a control, the dietary regimen was continued, the second group received a vitamin K-deficient diet, whereas the third group received the standardized control diet q one oral dose of 1 mg of brodifacoum per kg body weight. Blood and urine samples were taken at day 3 Žcontrols and brodifacoum group. and at day 7 Ž controls and vitamin K-deficient group.. The data are given in Table 3, and show that both after dietary vitamin K-deficiency as well as after brodifacoum treatment the circulating prothrombin concentrations fall to below 10% of the starting values. The urinary Gla-excretion, however, declined from 1.63 to 1.07 m molr day Žs 66%. in the vitamin K-deficient animals and from 1.59 to 0.30 m molrday Ž s 19%. in the brodifacoum-treated animals. From these data it appears that nutritional vitamin K-deficiency gave rise to a substantial decrease of prothrombin activity, which is
Fig. 2. Histogram for Gla-excretion in untreated rats. Rats were housed in individual metabolic cages, and all urine was collected during a period of 24 h. Both the means and the median for excretion were 1.5 m molr24 h.
only partly accompanied by a decrease in Gla-excretion. A complete blockade of the vitamin K cycle by coumarin derivatives resulted in very low prothrombin levels Žbelow the level of detection, i.e. - 2%. , and in a more pronounced reduction of Gla-excretion than in the case of vitamin K-deficiency. From the data in Table 3 it may be calculated that the ratio between urinary Gla-excretion and plasma prothrombin concentration was not constant, but increased 4.5-fold during 7 days of vitamin K-deficiency and even ) 8.2-fold after 3 days of brodifacoum treatment. This suggests that a substantial fraction of the urinary Gla originates from Gla-proteins with a much longer half-life time than prothrombin.
Table 3 Effect of vitamin K-deficiency and vitamin K-antagonists on plasma prothrombin and urinary Gla Control 1 n s 10
Brodifacoum-treatment n s 10
Control 2 n s 10
Vitamin K-deficiency n s 10
Gla-excretion Žnmolrday. Before treatment After treatment Difference Statistical significance Ž P .
1308 Ž"88. 1299 Ž"93. y9 Ž"34. N.S.
1587 Ž"141. 301 Ž"18. y1286 Ž"151. 0.005
1308 Ž"88. 1284 Ž"131. y24 Ž"18. N.S.
1625 Ž"59. 1065 Ž"74. y560 Ž"70. 0.007
Prothrombin Ž%. Before treatment After treatment Difference Statistical significance Ž P .
77.8 Ž"1.0. 81.5 Ž"2.2. q3.7 Ž"2.3. N.S.
86.3 Ž"3.9. -2 - y84.3 0.005
77.8 Ž"1.1. 80.0 Ž"1.7. q2.2 Ž"2.4. N.S.
72.3 Ž"1.5. 10.5 Ž"1.2. y61.8 Ž"2.0. 0.005
Statistical significance of differences was established with the Wilcoxon test. Furhter details are as in the legend to Table 1.
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
3.4. Correlation between the plasma prothrombin concentration and urinary Gla-excretion Although all experiments were performed using an inbred strain of rats, there was substantial interindividual variation in the plasma prothrombin concentration, which ranged from 70–110% Žmean value: 79%. of the pooled reference plasma. Urinary Gla-excretion ranged from 0.8–2.2 m molrday Žmean value: 1.5 m molrday, see Fig. 2.. Using the data from 30 animals which had been kept for 3 weeks on the standardized control diet no correlation was found between the plasma prothrombin level and the urinary Gla excretion Ž r s y0.17; P ) 0.4.. It was concluded that the interindividual differences in plasma prothrombin concentrations are not reflected in the urinary Gla excretion.
4. Discussion The experiments presented in this paper were designed to bring rats in a steady state situation of nutritional vitamin K intake; it was assumed that in this situation the amount of urinary Gla-excretion equalled the total body Gla synthesis. Furthermore it was assumed that the plasma prothrombin concentration was a reliable marker for hepatic Gla synthesis. By comparing these two variables at different nutritional vitamin K intake we were able to investigate the vitamin K-requirement for hepatic and extrahepatic Gla formation. The data in this paper showed an almost linear relation between the dietary vitamin K concentration and the plasma prothrombin between 0 and 6 m g of menaquinone-4 per g of food. Vitamin K-sufficiency is generally defined as the vitamin K intake at which no further increase of plasma prothrombin takes place. For our studies we have taken rats fed with 20 m g Ž45 nmol. of menaquinone-4 per g of food as vitamin K-sufficient control animals, and it was shown that between 45 and 2,000 nmol of K-vitamers per g of food there was no change in the prothrombin concentration. On the other hand, the excess of phylloquinone and menaquinone-4 induced a significant increase in total Gla-excretion, suggesting that the vitamin K-requirement of a number of extrahepatic gglutamylcarboxylases is higher than that of the hep-
49
atic enzyme. This is consistent with earlier results from human studies showing that a major fraction of the bone Gla-protein osteocalcin occurs in an undercarboxylated form w15x, and that supplementation of pharmacological doses of phylloquinone increases both the synthesis of osteocalcin, as well as its degree of carboxylation w19–21x. High doses of menaquinone-8, on the other hand, did not affect the urinary Gla-excretion. Groenen-van Dooren et al. show ed that the intestinal absorption of menaquinone-9 is slower than that of menaquinone-4 and phylloquinone, but these authors reported a maximal effect of menaquinone-9 at 24 h after ingestion of the vitamin. It must be assumed, therefore, that after 3 days of menaquinone-8 treatment the animals were in a steady state situation. This suggests that higher menaquinones do not have cofactor activity for extrahepatic carboxylases in vivo. Remarkably, the Gla-excretion had only slightly decreased in the vitamin K-deficient state. The most plausible explanation for this phenomenon is that during nutritional vitamin K-deficiency one or more sources of vitamin K remain available Ž e.g., liver or fat tissue stores, intestinal menaquinone production, coprophagy.. The fact that the animals generally started bleeding after 7–10 days demonstrates that these sources are insufficient to maintain normal haemostasis. On the other hand this residual vitamin K might be sufficient to produce coagulation factors Žor precursors thereof. with a reduced amount of Gla-residues. The occurrence of these undercarboxylated coagulation factors in the circulation has been demonstrated in coumarin-treated humans and cows w22,23x. Also if the undercarboxylated precursor proteins would be degraded intracellularly w24x, the Glaresidues cannot be metabolized further, and are bound to reach the circulation to be excreted via the urine. It is well known that undercarboxylated coagulation factors have little Žif any. procoagulant activity w25,26x, but they might still substantially contribute to the urinary Gla-excretion. In contrast to vitamin K-deficiency, brodifacoum efficiently and irreversibly blocks the vitamin Kcycle, thus preventing any further carboxylation events. Residual urinary Gla must originate from Gla-proteins formed before brodifacoum administration. After 3 days prothrombin Ž which is the coagulation factor with the longest half-life time, i.e., 9 h
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
50
w27x. had decreased below the level of detection, followed by serious bleeding during the fourth day. At that time the urinary Gla excretion had decreased to 17% of the starting value. Obviously the residual Gla originated from proteins with longer half-life times than prothrombin has, for instance those adsorbed to the hydroxyapatite matrix of bone.
Acknowledgements This research was supported by grant 92.307 from the Netherlands Heart Foundation.
References w1x w2x w3x w4x w5x w6x w7x w8x
w9x
Shearer, M.J. Ž1995. Lancet 345, 229–234. Vermeer, C. Ž1990. Biochem. J. 266, 625–636. Suttie, J.W. Ž1987. Hepatology 7, 367–376. Suttie, J.W. and Preusch, P.C. Ž1986. Haemostasis 16, 193– 215. Fasco, M.J., Hildebrandt, E.F. and Suttie, J.W. Ž1982. J. Biol. Chem. 257, 11210–11212. Mosterd, J.J. and Thijssen, H.H.W. Ž1991. Br. J. Pharmacol. 104, 531–535. Fernlund, P. Ž1976. Clin. Chim. Acta 72, 147–155. Lian, J.B., Glimacher, M.J. and Gallop, P.M. Ž1971. in Calcium Binding Proteins and Calcium Function ŽWasserman R.H. et al., eds.., pp. 379–381, Elsevier North Holland, Amsterdam. Shah, D.V., Tews, J.K., Harper, A.E. and Suttie, J.W. Ž1978. Biochim. Biophys. Acta 539, 209–217.
w10x Buchthal, S.D. and Bell, R.G. Ž1983. Biochemistry 22, 1077–1082. w11x Vermeer, C., Hendrix, H. and Daemen, M. Ž1982. FEBS Lett. 148, 317–320. w12x De Boer-van den Berg, M.A.G., Uitendaal, M.P. and Vermeer, C. Ž1987. Mol. Cell. Biochem. 5, 71–76. w13x Hauschka, P.V., Friedman, P.A., Traverso, H.P. and Gallop, P.M. Ž1976. Biophys. Res. Commun. 71, 1207–1213. w14x Hauschka, P.V., Lian, J.B., Cole, D.E.C. and Gundberg, C.M. Ž1989. Phys. Rev. 69, 990–1047. w15x Price, P.A. and Cairns, J.R. Ž1994. J. Bone Miner. Res. 9, 1989–1997. w16x Knapen, M.H.J., Jie, K.-S.G., Hamulyak, ´ K. and Vermeer, C. Ž1991. Thromb. Haemostas. 65, 671. w17x Groenen-van Dooren, M.M.C.L., Soute, B.A.M., Jie, K.S.G., Thijssen, H.H.W. and Vermeer, C. Ž1993. Biochem. Pharmacol. 433–437. w18x Kuwada, M. and Katayama, K. Ž1983. Analyt. Biochem. 131, 173–179. w19x Knapen, M.H.J., Hamulyak, ´ K. and Vermeer, C. Ž1989. Ann. Int. Med. 111, 1001–1005. w20x Knapen, M.H.J., Jie, K.-S.G., Hamulyak, ´ K. and Vermeer, C. Ž1993. Calcif. Tissue Int. 53, 81–85. w21x Plantalech, L.C., Chapuy, M.C., Guillaumont, M., Chapuy, P., Leclercq, M. and Delmas, P.D. Ž1990. in Osteoporosis 1990 ŽChristiansen, C. and Overgaard, K., eds.., pp. 345– 347, Osteopress, Copenhagen. w22x Friedman, P.A. Ž1984. N. Engl. J. Med. 310, 1458–1460. w23x Malhotra, O.P. Ž1987. Thromb. Res. 47, 493–500. w24x Stanton, C. and Wallin, R. Ž1992. Biochem. J. 284, 25–31. w25x Ratcliffe, J.V., Furie, B. and Furie, B.C. Ž1993. J. Biol. Chem. 268, 24339–24345. w26x Zhang, L. and Castellino, F.J. Ž1993. J. Biol. Chem. 268, 12040–12045. w27x Vainieri, H. and Wingard, L.B. Ž1977. J. Pharmacol. Exp.Ther. 201, 507–517.
Nutritional vitamin K-intake and urinary g-carboxyglutamate excretion in the rat Alexandra M. Craciun, Monique M.C.L. Groenen-van Dooren, Cees Vermeer
)
CardioÕascular Research Institute Maastricht and Department of Biochemistry, UniÕersity of Limburg, P.O. Box 616, 6200 MD Maastricht, The Netherlands Received 11 March 1996; revised 9 July 1996; accepted 17 July 1996
Abstract Using the rat as an experimental animal model we have found that prothrombin synthesis reaches its maximal level at a relatively low dietary vitamin K intake. At still higher vitamin K intakes, however, the urinary Gla-excretion was substantially increased, showing a different vitamin K requirement for liver and extrahepatic tissues. The increased urinary Gla-excretion was found for both phylloquinone and menaquinone-4, but not for menaquinone-8, which questions the bioavailability of higher menaquinones for extrahepatic tissues. A discrepancy was found between effects of nutritional vitamin K-deficiency and treatment with a vitamin K-antagonist Žbrodifacoum.. With both regimens plasma prothrombin rapidly decreased to well below 10% of the starting values, but in case of K-deficiency urinary Gla had hardly decreased in 7 days, whereas after 3 days of brodifacoum treatment Gla-excretion had decreased to 17% of the starting values. An explanation for this observation is that prothrombin procoagulant activity does not decrease proportional to the prothrombin Gla-content, but that a wide range of undercarboxylated prothrombins have lost nearly all activity. During vitamin K-deficiency the remaining low levels of vitamin K would mainly give rise to undercarboxylated prothrombin, whereas during brodifacoum treatment only non-carboxylated prothrombin is formed. It seems plausible that in the latter case the urinary Gla originates from proteins with long half-life times, such as the bone Gla-proteins. Keywords: Vitamin K; Phylloquinone; Menaquinone; g-Carboxyglutamate; Coagulation; Posttranslational
1. Introduction Vitamin K is a group name for a number of related compounds which all have a naphthoquinone ring structure, but which differ from each other by the length and degree of saturation of their aliphatic side
Abbreviations: Gla, g-carboxyglutamate Corresponding author. Fax: q31 43 3670992.
)
chain at the 3-position w1x. The three K-vitamers which were studied in this paper are: phylloquinone, menaquinone-4, and menaquinone-8. Phylloquinone and menaquinone-4 are different with respect to their degree of saturation, whereas menaquinone-4 and menaquinone-8 differ in their aliphatic side chain length. The only known function of vitamin K in mammals is to act as a coenzyme for the endoplasmic enzyme g-glutamylcarboxylase during the posttranslational conversion of protein-bound glutamate
0304-4165r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 0 4 - 4 1 6 5 Ž 9 6 . 0 0 0 7 3 - 6
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
residues into g-carboxyglutamate ŽGla. w2,3x. The daily requirement of vitamin K is extremely low, because a second enzyme ŽKO-reductase. is capable of recycling the vitamin so that one molecule may catalyze the formation of several thousand Gla-residues w2,4x. A number of coumarin derivatives Že.g., warfarin and brodifacoum. inhibit the recycling of vitamin K by blocking the KO-reductase, thus rapidly exhausting the pool of bioavailable vitamin K w5,6x. Once it has been formed Gla cannot be metabolized or re-used, so eventually all Gla is excreted stoichiometrically via the urine, most of it as a free amino acid and less than 5% in a peptide-bound form w7–9x. Hence urinary Gla levels may reflect the turnover of the total body pool of Gla proteins; in a steady state situation the urinary Gla excretion will even be a measure for the total Gla production. By comparing plasma prothrombin and urinary Gla under different conditions an impression may be obtained of the extent and vitamin K-requirement of extra-hepatic carboxylation events. The enzymes of the vitamin K-cycle Ž i.e., carboxylase and reductase. have been discovered in the liver, but afterwards they were detected in most tissues and cell types w10–13x. The hepatic system is involved in the production of several blood coagulation factors, the most abundant of which is prothrombin. Osteoblasts in bone produce osteocalcin and matrix Gla-protein, two proteins which are structurally not related to the hepatic Gla-proteins w14x. Recent studies in humans have indicated that the nutritional vitamin K-requirement for complete carboxylation of osteocalcin may be higher than that for the carboxylation of prothrombin and related coagulation factors w15,16x. In this paper we have made use of the rat as an experimental model, and we have assessed the minimal dietary vitamin K intake at which the production of the hepatic Gla-protein prothrombin takes place at its maximal rate. To exclude small variations due to interindividual differences all animals were adapted for 3 weeks to a diet containing 3 times the minimal vitamin K requirement. In this model we have investigated the effect on the plasma prothrombin concentration and the urinary Gla excretion of a very high intake of various K vitamers, that of a nutritional vitamin K-deficiency, and finally the effect of administration of the vitamin K-antagonist brodifacoum.
45
2. Materials and methods 2.1. Chemicals Phylloquinone Žvitamin K 1 ., D,L-g-carboxyglutamic acid ŽGla., and orthophthaldialdehyde ŽOPA. were purchased from Sigma ŽSt. Louis, MO.; menaquinone-4 and menaquinone-8 were kind gifts from Hoffmann-La Roche ŽBasel, Switzerland.. The non-ionic surfactant HCO-60 Žpolyoxyethylene hydrogenated castor oil derivatives. was purchased from Nikko Chemicals ŽTokyo, Japan.. Brodifacoum was a gift from Sorex ŽCheshire, UK. and it was dissolved to a final concentration of 0.25 mgrml in a buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.8 and 1% Žwrv. HCO-60 shortly before use. The various K-vitamers were dissolved in the same buffer system to a final concentration of 10 m molrml. All other chemicals were of analytical grade or better. 2.2. Animals and diets All studies were performed in male rats of the Lewis strain, which were 12 weeks old when entering in the experiment. The animals were housed in individual metabolic cages equipped for separate collection of urine and faeces ŽTecniplast 170102 cages, B and E Industry, Utrecht, The Netherlands., with a 12-h light-dark cycle and controlled temperature Ž 20 " 28C. and humidity Ž50 " 10%.. They had free access to food and water. The following foods were used: 1. A commercial rat diet Ž SRM-A, Hope Farms, Woerden, The Netherlands., the vitamin K content of which was found to be 5 mg of menadione and 2 mg of phylloquinone per kg of food. 2. A radiated Ž 0.9 Mrad. vitamin K-deficient diet ŽHope Farms, Woerden, The Netherlands., which was mixed with cooked and dried white rice in a 1:1 Žwrw. ratio prior to use w17x. This food was checked not to contain detectable amounts of vitamin K. 3. The same vitamin K-deficient diet, mixed with cooked and dried rice, to which 20 m g of menaquinone-4 per g of food were added. All food was powdered with a blender and mixed in a professional food processor before use. To ensure a controlled vitamin K intake the vitamers were administered orally, via a syringe equipped with a
46
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
plastic cannula, which was protruded into the esophagus. All rats were weighed regularly, and it was checked that their body weights increased by about 1 g per day, and that the various groups did not differ with respect to mean body weight gain. Blood Ž 0.5 ml. was collected in 0.05 ml of 0.14 M trisodium citrate after puncture of the tail vein. The blood samples were centrifuged for 15 min at 3 000 = g and the plasma was frozen at y808C until use. Urines were collected during the last 24 h of each measuring point, the volumes were recorded. Subsequently the samples were centrifuged for 10 min at 5 000 = g and the supernatants were frozen at y808C until use. 2.3. Urinary Gla detection The urine samples were diluted with distilled water between 4 and 160 times Žas required to stay within the normal detection range., and the diluted samples were analyzed according to the procedure of Kuwada and Katayama w18x, using a Separations high precision HPLC pump Žmodel 300., a Kratos Spectraflow 980 detector and a Nucleosil 100-5 SB anion exchange resin Ž5 m m particles, Macherey-Nagel, Duren, Germany. at a flow rate of 1 mlrmin. Refer¨ ence curves were prepared from samples with known concentrations of authentic Gla. The sensitivity of the assay was 0.3 m M. The intra- and interassay coefficients of variation were 2.2 and 6.4%, respectively.
and the Wilcoxon test for experiments with less animals per group. All data are expressed as mean values" S.E.
3. Results 3.1. Gla-excretion at the minimal Õitamin K intake required for normal blood coagulation Sixty rats were made vitamin K-deficient by feeding them with a commercial diet devoid of vitamin K for a period of 1 week. At that time blood was collected, and it was found that the circulating prothrombin concentration had decreased to 10–20% of the starting values. Subsequently the rats were subdivided into 10 groups, and varying Žknown. amounts of menaquinone-4 were added to the diet. This regime was continued for 2 weeks before blood was taken for another prothrombin determination. It was found that at a menaquinone-4 content of 6 m g Ž13.5 nmol. per g of food the prothrombin levels had reached their pre-treatment levels, and that they did not change at higher intakes Žsee Fig. 1.. Since the mean food consumption was 22 g per day, it may be calculated that the minimal dietary menaquinone-4 requirement was around 300 nmol per day. On the basis of these data it was decided that for all experiments described below the reference rats Ži.e., vitamin K-sufficient. would receive a diet containing 20 m g Ž45 nmol. of
2.4. Plasma prothrombin assay Prothrombin concentrations were determined in the one-stage assay using a coagulometer Ž KC-4, Amelung, Germany. , and a commercial thromboplastin preparation ŽThromborel S w . and clotting factor II deficient plasma Ž both from Behringwerke AG, Marburg, Germany.. Prothrombin concentrations were calculated with the aid of reference curve from pooled normal rat plasma. 2.5. Statistical analysis The software package SPSSrPCq version 5.0 ŽSPSS, Chicago, IL. was used for the statistical analysis and the 0.05 level of significance was chosen. The data were analyzed using the paired Student’s t-test for comparing groups of more than 20 animals,
Fig. 1. Menaquinone-4 requirement for normal prothrombin synthesis in the vitamin K-deficient rat. The food was administered ad libitum during a period of 2 weeks. Plasma prothrombin concentrations are expressed as a percentage of the starting values. Each point represents the mean value Ž"S.E.. for six rats.
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
47
Table 1 Effect of diet on plasma prothrombin and urinary Gla excretion Marker
Before treatment
After treatment
Difference
P-value
Prothrombin Ž%. Gla-excretion Žnmolrday. Glarprothrombin ratio
80.2 Ž"0.8. 1470 Ž"70. 18.4 Ž"1.
79.9 Ž"1.8. 1340 Ž"70. 17.0 Ž"1.
y0.28 Ž"1.7. y130 Ž"40. y1.4 Ž"0.5.
N.S. - 0.005 - 0.02
Measurements were made in rats before and after feeding for three weeks with the standardized control diet. Prothrombin was expressed relative to a reference pool of rat plasma, Gla-excretion was measured in 24-h urines. The significance of differences was calculated with the paired Student’s t-test. N.S. stands for not significant Ž P ) 0.05..
menaquinone-4 per g. This diet will be referred to as the ‘standardized control diet’. Reference rats were defined further in an experiment in which we have measured the plasma prothrombin concentration and the urinary Gla excretion in 25 animals before they entered the experiment and after they had been fed for 3 weeks with the standardized control diet. The data are given in Table 1, and demonstrate that plasma prothrombin had not changed by the switch from commercial food to the standardized control diet used for known vitamin K-intake. The urinary Gla-excretion had slightly decreased, however, which suggests that the Gla-excretion can be decreased without changing the plasma prothrombin concentration, and also that more vitamin K can be extracted from the commercial chow than from the standardized control diet. The divergence between urinary Gla excretion and plasma prothrombin concentration is visualized most clearly from the ratio between both variables. As can be seen from Table 1
this ratio had decreased significantly after the dietary change. 3.2. Effects of high doses of Õarious K-Õitamers on plasma prothrombin and on urinary Gla Thirty five rats were kept for 3 weeks on the standardized control diet, and subsequently they were subdivided into three groups of 10 and one of group of 5 animals. The first group Ž n s 10. continued with the standardized control diet as its sole source of nutritional vitamin K, the other three groups received a high oral dose Ž20 m moles. of vitamin K on each of three successive days. Three vitamers were studied: phylloquinone Ž n s 10., menaquinone-4 Ž n s 10., and menaquinone-8 Ž n s 5.. As is shown in Table 2, these high doses of vitamin K did not change the circulating prothrombin concentrations. On the other hand, urinary Gla excretion was substantially and significantly increased by both phylloquinone and
Table 2 Effect of high vitamin K intake on plasma prothrombin and urinary Gla Control n s 10
Excess K 1 n s 10
Excess MK-4 n s 10
Excess MK-8 ns5
Gla-excretion Žnmolrday. Before treatment After treatment Difference Statistical significance Ž P .
1464 Ž"178. 1451 Ž"165. y13 Ž"50. N.S.
1501 Ž"204. 1853 Ž"232. q352 Ž"110. 0.02
1372 Ž"280. 1875 Ž"406. q503 Ž"210. 0.01
1330 Ž"75. 1486 Ž"86. q156 Ž"70. N.S.
Prothrombin Ž%. Before treatment After treatment Difference Statistical significance Ž P .
77.2 Ž"1.8. 83.4 Ž"3.4. q6.2 Ž"7.5. N.S.
84.4 Ž"3.7. 89.9 Ž"5.1. q5.5 Ž"7.1. N.S.
85.9 Ž"4.5. 90.6 Ž"6.4. q4.7 Ž"8.3. N.S.
81.6 Ž"2.1. 89.2 Ž"3.8. q7.6 Ž"3.5. N.S.
K 1 stands for phylloquinone, MK for menaquinone. Statistical significance of differences was calculated with the Wilcoxon test. Further details are as in the legend to Table 1.
48
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
menaquinone-4, but not by menaquinone-8. Prolonged treatment with the high doses of phylloquinone and menaquinone-4 until day 7 did not change these data further Ždata not shown.. Unfortunately our supply was insufficient to continue the treatment with menaquinone-8. 3.3. Effects of dietary Õitamin K-deficiency and oral anticoagulants on plasma prothrombin and on urinary Gla Thirty rats were kept for 3 weeks on the standardized control diet, and subsequently they were subdivided into three groups of 10 animals each. In the first group, which served as a control, the dietary regimen was continued, the second group received a vitamin K-deficient diet, whereas the third group received the standardized control diet q one oral dose of 1 mg of brodifacoum per kg body weight. Blood and urine samples were taken at day 3 Žcontrols and brodifacoum group. and at day 7 Ž controls and vitamin K-deficient group.. The data are given in Table 3, and show that both after dietary vitamin K-deficiency as well as after brodifacoum treatment the circulating prothrombin concentrations fall to below 10% of the starting values. The urinary Gla-excretion, however, declined from 1.63 to 1.07 m molr day Žs 66%. in the vitamin K-deficient animals and from 1.59 to 0.30 m molrday Ž s 19%. in the brodifacoum-treated animals. From these data it appears that nutritional vitamin K-deficiency gave rise to a substantial decrease of prothrombin activity, which is
Fig. 2. Histogram for Gla-excretion in untreated rats. Rats were housed in individual metabolic cages, and all urine was collected during a period of 24 h. Both the means and the median for excretion were 1.5 m molr24 h.
only partly accompanied by a decrease in Gla-excretion. A complete blockade of the vitamin K cycle by coumarin derivatives resulted in very low prothrombin levels Žbelow the level of detection, i.e. - 2%. , and in a more pronounced reduction of Gla-excretion than in the case of vitamin K-deficiency. From the data in Table 3 it may be calculated that the ratio between urinary Gla-excretion and plasma prothrombin concentration was not constant, but increased 4.5-fold during 7 days of vitamin K-deficiency and even ) 8.2-fold after 3 days of brodifacoum treatment. This suggests that a substantial fraction of the urinary Gla originates from Gla-proteins with a much longer half-life time than prothrombin.
Table 3 Effect of vitamin K-deficiency and vitamin K-antagonists on plasma prothrombin and urinary Gla Control 1 n s 10
Brodifacoum-treatment n s 10
Control 2 n s 10
Vitamin K-deficiency n s 10
Gla-excretion Žnmolrday. Before treatment After treatment Difference Statistical significance Ž P .
1308 Ž"88. 1299 Ž"93. y9 Ž"34. N.S.
1587 Ž"141. 301 Ž"18. y1286 Ž"151. 0.005
1308 Ž"88. 1284 Ž"131. y24 Ž"18. N.S.
1625 Ž"59. 1065 Ž"74. y560 Ž"70. 0.007
Prothrombin Ž%. Before treatment After treatment Difference Statistical significance Ž P .
77.8 Ž"1.0. 81.5 Ž"2.2. q3.7 Ž"2.3. N.S.
86.3 Ž"3.9. -2 - y84.3 0.005
77.8 Ž"1.1. 80.0 Ž"1.7. q2.2 Ž"2.4. N.S.
72.3 Ž"1.5. 10.5 Ž"1.2. y61.8 Ž"2.0. 0.005
Statistical significance of differences was established with the Wilcoxon test. Furhter details are as in the legend to Table 1.
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
3.4. Correlation between the plasma prothrombin concentration and urinary Gla-excretion Although all experiments were performed using an inbred strain of rats, there was substantial interindividual variation in the plasma prothrombin concentration, which ranged from 70–110% Žmean value: 79%. of the pooled reference plasma. Urinary Gla-excretion ranged from 0.8–2.2 m molrday Žmean value: 1.5 m molrday, see Fig. 2.. Using the data from 30 animals which had been kept for 3 weeks on the standardized control diet no correlation was found between the plasma prothrombin level and the urinary Gla excretion Ž r s y0.17; P ) 0.4.. It was concluded that the interindividual differences in plasma prothrombin concentrations are not reflected in the urinary Gla excretion.
4. Discussion The experiments presented in this paper were designed to bring rats in a steady state situation of nutritional vitamin K intake; it was assumed that in this situation the amount of urinary Gla-excretion equalled the total body Gla synthesis. Furthermore it was assumed that the plasma prothrombin concentration was a reliable marker for hepatic Gla synthesis. By comparing these two variables at different nutritional vitamin K intake we were able to investigate the vitamin K-requirement for hepatic and extrahepatic Gla formation. The data in this paper showed an almost linear relation between the dietary vitamin K concentration and the plasma prothrombin between 0 and 6 m g of menaquinone-4 per g of food. Vitamin K-sufficiency is generally defined as the vitamin K intake at which no further increase of plasma prothrombin takes place. For our studies we have taken rats fed with 20 m g Ž45 nmol. of menaquinone-4 per g of food as vitamin K-sufficient control animals, and it was shown that between 45 and 2,000 nmol of K-vitamers per g of food there was no change in the prothrombin concentration. On the other hand, the excess of phylloquinone and menaquinone-4 induced a significant increase in total Gla-excretion, suggesting that the vitamin K-requirement of a number of extrahepatic gglutamylcarboxylases is higher than that of the hep-
49
atic enzyme. This is consistent with earlier results from human studies showing that a major fraction of the bone Gla-protein osteocalcin occurs in an undercarboxylated form w15x, and that supplementation of pharmacological doses of phylloquinone increases both the synthesis of osteocalcin, as well as its degree of carboxylation w19–21x. High doses of menaquinone-8, on the other hand, did not affect the urinary Gla-excretion. Groenen-van Dooren et al. show ed that the intestinal absorption of menaquinone-9 is slower than that of menaquinone-4 and phylloquinone, but these authors reported a maximal effect of menaquinone-9 at 24 h after ingestion of the vitamin. It must be assumed, therefore, that after 3 days of menaquinone-8 treatment the animals were in a steady state situation. This suggests that higher menaquinones do not have cofactor activity for extrahepatic carboxylases in vivo. Remarkably, the Gla-excretion had only slightly decreased in the vitamin K-deficient state. The most plausible explanation for this phenomenon is that during nutritional vitamin K-deficiency one or more sources of vitamin K remain available Ž e.g., liver or fat tissue stores, intestinal menaquinone production, coprophagy.. The fact that the animals generally started bleeding after 7–10 days demonstrates that these sources are insufficient to maintain normal haemostasis. On the other hand this residual vitamin K might be sufficient to produce coagulation factors Žor precursors thereof. with a reduced amount of Gla-residues. The occurrence of these undercarboxylated coagulation factors in the circulation has been demonstrated in coumarin-treated humans and cows w22,23x. Also if the undercarboxylated precursor proteins would be degraded intracellularly w24x, the Glaresidues cannot be metabolized further, and are bound to reach the circulation to be excreted via the urine. It is well known that undercarboxylated coagulation factors have little Žif any. procoagulant activity w25,26x, but they might still substantially contribute to the urinary Gla-excretion. In contrast to vitamin K-deficiency, brodifacoum efficiently and irreversibly blocks the vitamin Kcycle, thus preventing any further carboxylation events. Residual urinary Gla must originate from Gla-proteins formed before brodifacoum administration. After 3 days prothrombin Ž which is the coagulation factor with the longest half-life time, i.e., 9 h
A.M. Craciun et al.r Biochimica et Biophysica Acta 1334 (1997) 44–50
50
w27x. had decreased below the level of detection, followed by serious bleeding during the fourth day. At that time the urinary Gla excretion had decreased to 17% of the starting value. Obviously the residual Gla originated from proteins with longer half-life times than prothrombin has, for instance those adsorbed to the hydroxyapatite matrix of bone.
Acknowledgements This research was supported by grant 92.307 from the Netherlands Heart Foundation.
References w1x w2x w3x w4x w5x w6x w7x w8x
w9x
Shearer, M.J. Ž1995. Lancet 345, 229–234. Vermeer, C. Ž1990. Biochem. J. 266, 625–636. Suttie, J.W. Ž1987. Hepatology 7, 367–376. Suttie, J.W. and Preusch, P.C. Ž1986. Haemostasis 16, 193– 215. Fasco, M.J., Hildebrandt, E.F. and Suttie, J.W. Ž1982. J. Biol. Chem. 257, 11210–11212. Mosterd, J.J. and Thijssen, H.H.W. Ž1991. Br. J. Pharmacol. 104, 531–535. Fernlund, P. Ž1976. Clin. Chim. Acta 72, 147–155. Lian, J.B., Glimacher, M.J. and Gallop, P.M. Ž1971. in Calcium Binding Proteins and Calcium Function ŽWasserman R.H. et al., eds.., pp. 379–381, Elsevier North Holland, Amsterdam. Shah, D.V., Tews, J.K., Harper, A.E. and Suttie, J.W. Ž1978. Biochim. Biophys. Acta 539, 209–217.
w10x Buchthal, S.D. and Bell, R.G. Ž1983. Biochemistry 22, 1077–1082. w11x Vermeer, C., Hendrix, H. and Daemen, M. Ž1982. FEBS Lett. 148, 317–320. w12x De Boer-van den Berg, M.A.G., Uitendaal, M.P. and Vermeer, C. Ž1987. Mol. Cell. Biochem. 5, 71–76. w13x Hauschka, P.V., Friedman, P.A., Traverso, H.P. and Gallop, P.M. Ž1976. Biophys. Res. Commun. 71, 1207–1213. w14x Hauschka, P.V., Lian, J.B., Cole, D.E.C. and Gundberg, C.M. Ž1989. Phys. Rev. 69, 990–1047. w15x Price, P.A. and Cairns, J.R. Ž1994. J. Bone Miner. Res. 9, 1989–1997. w16x Knapen, M.H.J., Jie, K.-S.G., Hamulyak, ´ K. and Vermeer, C. Ž1991. Thromb. Haemostas. 65, 671. w17x Groenen-van Dooren, M.M.C.L., Soute, B.A.M., Jie, K.S.G., Thijssen, H.H.W. and Vermeer, C. Ž1993. Biochem. Pharmacol. 433–437. w18x Kuwada, M. and Katayama, K. Ž1983. Analyt. Biochem. 131, 173–179. w19x Knapen, M.H.J., Hamulyak, ´ K. and Vermeer, C. Ž1989. Ann. Int. Med. 111, 1001–1005. w20x Knapen, M.H.J., Jie, K.-S.G., Hamulyak, ´ K. and Vermeer, C. Ž1993. Calcif. Tissue Int. 53, 81–85. w21x Plantalech, L.C., Chapuy, M.C., Guillaumont, M., Chapuy, P., Leclercq, M. and Delmas, P.D. Ž1990. in Osteoporosis 1990 ŽChristiansen, C. and Overgaard, K., eds.., pp. 345– 347, Osteopress, Copenhagen. w22x Friedman, P.A. Ž1984. N. Engl. J. Med. 310, 1458–1460. w23x Malhotra, O.P. Ž1987. Thromb. Res. 47, 493–500. w24x Stanton, C. and Wallin, R. Ž1992. Biochem. J. 284, 25–31. w25x Ratcliffe, J.V., Furie, B. and Furie, B.C. Ž1993. J. Biol. Chem. 268, 24339–24345. w26x Zhang, L. and Castellino, F.J. Ž1993. J. Biol. Chem. 268, 12040–12045. w27x Vainieri, H. and Wingard, L.B. Ž1977. J. Pharmacol. Exp.Ther. 201, 507–517.