104212 Clin Pathol 1993;46:1042-1045 Quantitation of in serum: Evaluation of its usefulness in routine clinical practice J Clin Pathol: first published as 10.1136/jcp.46.11.1042 on 1 November 1993. Downloaded from

N A M Boyd, A R Bradwell, RA Thompson

Abstract The main organ of synthesis of vitronectin Aims-To make a preliminary assess- is the liver,'9 but it is also produced by ment of the clinical relevance of serum platelets,20 monocytes, macrophages,21 and vitronectin concentrations in various dis- megakaryocytes.22 It exists as two polypep- ease groups, using a recently available tides of molecular weight 75 000 kilodaltons commercial radial immunodiffusion kit. (one chain form) and 65 000 kilodaltons Methods-Serum vitronectin concentra- (two chain form) in plasma and serum. The tions were measured in 80 control sub- 75 000 kilodalton precursor yields the 65 000 jects and 144 patients with various by proteolytic cleavage at the COOH diseases. The following characteristics terminal.2' Individuals vary in the ratios of the were used to evaluate the test proce- one and two chain forms.24 Previous analysis dures: linearity of method, inter- and of normal subjects and patients with throm- intrabatch precision, effect of storage, botic disorders showed three patterns: mostly temperature and in vitro activation of two chain; equal amounts of both forms; and the classical and alternative complement mostly one chain. Patterns are an inherited pathways on vitronectin concentrations. trait and are present in the Caucasian popula- Results-Significantly reduced serum tion in the ratios predicted by the Hardy vitronectin concentrations were found in Weinberg equilibrium of two alleles.25 The patients with liver disease, renal disease, gene frequency for the trait varies in different and systemic lupus erythematosus (SLE) populations. Caucasians have a lower fre- (normal C3 and C4 concentrations, when quency for the one chain form than Chinese26 compared with normal subjects. This and Japanese.27 particular method was suitable for mea- Serum vitronectin concentrations do not suring vitronectin concentrations in differ significantly from plasma concentra- serum samples provided they were tions.28 A wide range of plasma concentra-

stored at -20°C. tions for vitronectin have been reported http://jcp.bmj.com/ Conclusions-The clinical value of mea- between 0-14 mg/l1 and 0-6 mg/1.29 The suring serum vitronectin seems to be mean concentration measured in plasma in limited, but a larger study may be justi- healthy subjects has been reported as 0 5 fied to ascertain the clinical importance mg/1.30 There is no significant difference of reduced serum vitronectin concentra- between adult concentrations of plasma vit- tions in liver diseases, and the possible ronectin in men and women and there is no

role of vitronectin in other disease significant change in the vitronectin concen- on September 24, 2021 by guest. Protected copyright. processes. tration with age.4 The blood concentration measured in healthy newborns, however, is (7 Clin Pathol 1993;46:1042-1045) 67% of the adult level."3 Vitronectin plasma concentrations have been measured by rocket immunoelec- Vitronectin, also known as serum spreading trophoresis in patients with chronic liver dis- factor,' S-,2 and epibolin,' is a multi- ease; significantly decreased concentrations functional present in blood, were found in patients with liver cirrhosis and urine, amniotic fluid4 and the extracellular these changes correlated closely with changes Regional Immunology matrix of tissue.5 It has a role in the regula- in serum cholinesterase, factor X, and C3.'9 It Laboratory, East tion of both the complement6 and coagulation has been suggested recently that plasma con- Birmingham Hospital, Bordesley Green East, systems.78 Vitronectin binds membrane centrations of vitronectin may be useful as a Birmingham B9 SST bound C5b-99 and non-membrane bound parameter of hepatic synthetic function and N A M Boyd C5b-910 and it is known to inhibit cell lysis in as a marker of the severity of cirrhosis.'2 R A Thompson vitro." It also binds to heparin and neu- Substantially reduced plasma vitronectin con- The Medical School, tralises the inactivation of thrombin by centrations, measured by ELISA, have been University of Birmingham antithrombin III.12 Vitronectin has growth reported in patients with disseminated AR Bradwell promoting activity,"3 has recently been intravascular coagulation, especially if liver Correspondence to: described as an opsonin,'4 and is a cell adhe- failure is also present. Normal vitronectin Dr N A M Boyd, Regional Immunology Laboratory, sion molecule.'5 The vitronectin receptor concentrations have been detected in patients Belfast City Hospital, belongs to the beta3 integrin family which with acute leukaemia and metastatic cancer.25 Lisburn Road, Belfast BT9 7AD. recognise the peptide sequence Arg-Gly-Asp There is as yet no clear understanding of the Accepted for publication (RSD) on cells.'6 Vitronectin is a precursor of importance of these variations in the concen- 1 1 June 1993 the serum peptide somatomedin B.17 18 trations of vitronectin in different diseases. Quantitation ofvitronectin in serum 1043

Methods segmental glomerulosclerosis, four with mini- In evaluating the kit the following features mal change disease and four with IgA were examined: linearity of method; inter- nephropathy. The patients with SLE were J Clin Pathol: first published as 10.1136/jcp.46.11.1042 on 1 November 1993. Downloaded from and intrabatch precision; sample type; effect referred from rheumatologists and nephrolo- of storage; effect of temperature; effect of in gists in the West Midlands. All patients with vitro activation of the classical and alternative SLE had antibodies to double stranded complement pathway. DNA: they were divided into two groups, Two control samples and the calibrator those with normal C3 and C4 concentrations provided with the kit were tested neat and at and those with low C3 and C4 concentra- 1 in 2, 1 in 4, 1 in 8, 1 in 16 dilutions in tions. Patients with seropositive rheumatoid physiological saline. The neat sample was arthritis were referred from the rheumatolo- regarded as the true value. gists at East Birmingham Hospital and the For intrabatch analysis, one control serum patients with primary hypogammaglobuli- sample was assayed neat and at a 1 in 2 dilu- naemia attended the Regional Immunology tion five times using the same kit. For inter- Department, East Birmingham Hospital. batch analysis, one control serum sample was Four members from a family with C7 defi- assayed on ten separate occasions. The coeffi- ciency were also studied. One was homozy- cient of variation (CV) was calculated for gous C7 deficient and the others were each. heterozygous for the deficiency. All were clin- Serum and plasma (EDTA) samples on ically well at the time blood was taken. eight control subjects (four female and four Serum samples obtained from control sub- male) were tested. jects and patients were separated, aliquoted, Two control serum samples were aliquoted and stored at -20°C before testing. Vitro- and stored at 370C, 4°C, room temperature, nectin concentrations were measured using a -20°C and -700C for five days before testing. single radial immunodiffusion assay (The Two control samples were heated to 560C Binding Site, Birmingham, England). A cali- before testing. brator was provided with the kit and serum For classical complement pathway activa- samples were tested neat. The radial immuno- tion, 0-5 ml aggregated IgG was added to an diffusion plates were read after 48 hours. C3 equal volume of one control serum sample and C4 concentrations were measured by and incubated at 37°C for 30 minutes before nephelometry (Behring): SPS-0 1 (Supra testing. For alternative complement pathway Regional Specific Protein Unit, Sheffield) was activation, 05 ml zymosan was added to 0 5 used as the serum calibrant. ml of the same control serum and incubated Statistical analysis was performed using the at 37°C for 30 minutes before testing. Mann-Whitney U test and the Spearman The study population comprised a total of rank correlation coefficient test. 80 healthy blood donors who were used as the control group (40 women and 40 men,

age range 20-64) and 144 patients who were Results http://jcp.bmj.com/ divided into eight disease groups. Details are The first control sample gave a percentage summarised in table 1. recovery of 100% at 1 in 2, 1 in 4, and 1 in 8 Patients with liver disease were referred dilutions. The second control sample gave a from the Liver Unit at the Queen Elizabeth percentage recovery of 82% at the 1 in 2 dilu- Hospital, Birmingham. They included six tion and 100% at the 1 in 4 and 1 in 8 dilu- patients with primary biliary cirrhosis and tions. It was impossible to read the ring

four with chronic active hepatitis. There diameter of the 1 in 16 dilution. The control on September 24, 2021 by guest. Protected copyright. were single cases of a,antitrypsin deficiency, serum samples behaved in a similar manner haemachromatosis, non-A non-B hepatitis, to the calibrator. 3,4-methylenedioxymethamphetamine ("ecstasy") Intrabatch analysis CV for the control induced jaundice, fulminant hepatitis, possi- sample, when tested neat, was 1-99% and bly induced by a non-steroidal inflammatory 3*32% at the 1 in 2 dilution. The interbatch drug, hepatoma, alcholic cirrhosis, together analysis CV for the control serum sample was with a 6 year old boy who had been trans- 4.5%. planted for cryptogenic cirrhosis. Renal There was no significant difference ob- patients were referred from the Renal Unit at tained between measuring vitronectin con- the Queen Elizabeth Hospital, Birmingham. centrations in serum and plasma (EDTA) They included five patients with membranous samples. glomerulonephritis, two with mesangiopro- There was no significant difference be- liferative glomerulonephritis, three with focal tween serum vitronectin concentrations of samples stored at -200C and -700C. Vitro- nectin concentrations in both serum samples Table 1 stored at 37°C, 40C, and room temperature Groups tested Males Females Age Range were reduced (18-24%) when compared with Controls 40 40 20-64 samples stored at -200C. Iiver disease 10 10 6-76 Serum vitronectin concentrations mea- Renal disease 14 6 22-72 SLE (normal C3 and C4) 3 17 19-66 sured in control samples heated to 56°C were SLE (low C3 and C4) 1 19 6-56 reduced (23-27%) when compared with Rheumatoid arthritis 2 18 12-27 Primary hypogamma- 16 4 11-72 samples stored at -20°C. globulinaemia The serum vitronectin concentration was C7 deficiency 2 2 17-48 decreased by 51% following activation of the 1044 Boyd, Bradwell, Thompson

Table 2 was not strictly linear, and as with a number Mean of other plasma , tended to be geo- vitronectin Mean Mean metric. There was no significant difference J Clin Pathol: first published as 10.1136/jcp.46.11.1042 on 1 November 1993. Downloaded from Group concentratin C3 concentration C4 concentration Tested (range mgil) (range gil) (range gil) between adult concentrations of serum vit- ronectin in men and women, and there was Controls 475 1 19 0-31 (239-711) (0-75-1-63) (0-15-0-47) no significant change in concentrations with Liver disease 322 1-02 0-25 age. The mean concentrations and ranges of (106-538) (0-241-8) (0 01-0-49) Renal disease 405 0 99 0-41 vitronectin, C3, and C4 calculated in the (185-625) (0-55-1-43) (0 2-0 6) control subjects and patient groups are given SLE 413 1.0 0-26 (normal C3 and C4) (207-619) (0-48-1-52) (0-12-0-4) in table 2. SLE 405 0 59 0 11 The concentration of vitronectin in serum (low C3 and C4) ( 81-729) (0-31-0-87) (0 01-0 21) Rheumatoid 448 1-36 0-34 from patients with liver disease (p < 0-001), arthritis (295-601) (0-85-1-86) (0-16-0-52) renal disease (p = 0 018), and SLE with nor- Hypogamma- 455 1-37 0-4 globulinaemia (219-691) (0-81-1-93) (0-12-0-68) mal C3 and C4 concentrations (p < 0'022) C7 deficiency 524 1-04 0 3 was significantly lower than in control sub- (243-806) (0-75-1-33) (0-16-0-44) jects (fig IA). There was no significant difference in the concentration of vitronectin in serum from alternative complement pathway. Activation patients with SLE with low C3 and C4 con- of the classical complement pathway did not centrations, primary hypogammaglobuli- result in a decrease in vitronectin. naemia, rheumatoid arthritis, and C7 The mean concentration of vitronectin in deficiency when compared with control sub- serum from control subjects was 475 mg/l jects (fig 1A and B). There was a significant with quite a wide range (239-71 1). The dis- correlation between the serum vitronectin tribution of concentrations in the controls concentration in patients with liver disease and C4 concentrations, RS = 0 57 (p = 800- A 0008).

700- E Discussion . Vitronectin concentrations in plasma have c 600 *I I co0 *-U- already been measured by rocket immuno- X 500- * and ELISA aI. . electrophoresis'9 techniques.25 We a evaluated a radial immunodiffusion method 8 400- *li I. a; . a for the measurement of vitronectin. Our find- 0 . I 0 a a ings indicate that this commercially available .E 300 kit (The Binding Site, Birmingham, England) 0) is O 200- suitable for measuring vitronectin concen-

trations in serum samples stored at -20GC. http://jcp.bmj.com/ > 100- While the precision of the method was satis- factory, it was not directly compared with any 0 other method, because these are not yet com- Controls Liver Renal SLE - SLE - mercially available. Vitronectin was not stable normal low (C3 and C4) (C3 and C4) at 37'C, room temperature, or 4°C, and was denatured by heating to 56°C for 30 minutes. Figure JA Serum vitronectin concentrations in control subjects andpatients with liver disease, renal disease, SLE with normal C3 and C4 concentrations and SLE with low C3 Activation of the alternative complement on September 24, 2021 by guest. Protected copyright. and C4 concentrations. The horizontal lines indicate mean ISD. pathway by zymosan resulted in a decreased vitronectin concentration, but activation by aggregated IgG did not result in any signifi- cant change. 800- B Serum vitronectin concentrations were measured in control subjects and patients CD 700- with various diseases. The mean concentra- E tion in control subjects was 475 mg/l (range c 600- . I . 0 239-711). The present study confirmed pre-

+u a U.I CD I I~~~~~ vious reports that serum vitronectin concen- ' 500- I. trations in adult men and women do not ... 0) -*- C 400 .... differ and that there is no significant change 0 I; with age.4 It also confirmed that there is no c *a.)r-300-00 significant difference between vitronectin 0) concentrations measured in serum and o 200- plasma (EDTA) samples.28 There was a significant reduction in the 100- serum vitronectin concentration in our group of patients with liver disease, but there 1l n were insufficient cases of any one condition i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Control Hypogamma- Rheumatoid C7 Deficiency globulinaemia arthritis to permit correlation with the vitronectin concentration. In the group of patients with Figure lB Serum vitronectin concentration in control subjects andpatients with primary hypogammaglobulinaemia, rheumatoid arthritis, and C7 deficiency. Horizontal lines liver disease the C4 concentration correlated indicate mean ISD. with the serum vitronectin concentration. Quantitation ofvirronectin in serum 1045

Decreased plasma vitronectin concentrations protein, an inhibitor of the membrane attack complex of complement. J. Biol Chem 1979;254:9908-14. have been found before in patients with 7 Podack ER, Dahlback B, Griffin JH. Interaction of S-

chronic liver disease, and significantly protein of complement with thrombin and antithrombin J Clin Pathol: first published as 10.1136/jcp.46.11.1042 on 1 November 1993. Downloaded from m during coagulation. JBiol Chem 1986;261:7387-92. reduced concentrations in patients with liver 8 Preissner KT, Wassmuth R, Muler-Berghaus G. cirrhosis. In the present study significantly Physiochemical characterization of human S-protein and its function in the blood coagulation system. reduced serum concentrations of vitronectin BiochemJ3 1985;231:349-55. were also found in patients with renal disease 9 Bhakdi S, Kaflein R, Hastensen TS, Hugo F, Preissner KT, Mollnes TE. Complement S-protein (vitronectin) and patients with SLE with normal C3 is associated with cytolytic membrane-bound C5b-9 and C4 concentrations. The numbers of complexes. Clin Exp Immunol 1988;74:459-64. 10 Bhakdi S, Hugo F, Tranum-Jensen J. Functions and rele- patients in these groups, however, were again vance of the terminal complement sequence. Blut 1990; small. There was no significant difference in 60:309-18. 11 Dahlback B, Podack ER. Characterization of human S- the serum vitronectin concentration in the protein, an inhibitor of the membrane attack complex of groups of patients with SLE with reduced C3 complement. Demonstration of a free reactive thiol group. Biochemistry 1985;24:2368-74. and C4 concentrations, rheumatoid arthritis, 12 Preissner KT, Muller-Berghaus G. S-protein modulates or primary hypogammaglobulinaemia. the heparin-catalyzed inhibition of thrombin by antithrombin III. EurJ'Biochemz 1986;156:645-50. Decreased C3 and C4 concentrations in 13 Barnes DW, Vander Bosch J, Mujazaki K, Sato G. The some patients with SLE suggest active disease culture of human tumour cells. Methods Enzymol 1981;79:368-91. with activation of the classical complement 14 Anonymous [editorial]. Fibronectins and vitronectin. pathway. It is interesting that in vitro activa- Lancet 1989;i:474-6. 15 Pytela R, Piershbacher M, Ruoslahti E. A 125/115 KDa tion of the classical complement pathway did cell surface receptor specific for vitronectin interacts not produce a reduction in the vitronectin with the arginine-glycine-aspartic acid adhesion sequence derived from fibronectin. Proc Nail Acad Sci concentration either. While it is difficult to USA 1985;82:5766-70. explain why patients with SLE with normal 16 Ruoslahti E. Fibronectin and its receptors. Ann Rev Biochem 1988;57:375-413. C3 and C4 concentrations had significantly 17 Fryklund L, Sievertsson H. Primary structure of low vitronectin concentrations, this may be somatomedin B. FEBS Letts 1978;87:55-60. 18 Heldin C, Wasteson A, Fryklund L, Westermark B. because of alternative pathway activation in Somatomedin B: mitogenic activity derived from conta- these patients or some non-immune complex minant epidermal growth factor. Science 1981;213: 1122-3. mediated mechanism. 19 Kemkes-Matthes B, Preissner KT, Langenscheidt F, Inherited deficiencies of most regulatory Matthes KJ, Muller-Berghaus G. S-protein/vitronectin in chronic liver diseases: correlations with serum complement proteins have been found in cholinesterase, coagulation factor X and complement humans and disease associations are well components C3. EurJ Haematol 1987;39:161-5. 20 Preissner KT, Holzhuter S, Muller-Berghaus G. recognised. But no genetic deficiency of vit- Identification and partial characterization of platelet S- ronectin has been reported hitherto. protein. Haemostasis 1988;18:149. 21 Hetland G, Pettersen HB, Mollness TE, Johnson E. S- While a reduction in vitronectin concentra- protein is synthesized by human monocytes and tions were found in the groups mentioned macrophages in vitro. Scand Jr Immunol 1989;29: 15-21. 22 Kanz L, Lohr GW, Preissner KT. Identification of human above, the changes were less striking than megakaryocyte vitronectin/S-protein. Blood 1988;72: those in other routinely measured parameters, (Suppl):372a. 23 Hayman EG, Engvall E, A'Heam E, Barnes D, and it seems that this test is only a modest Pierschbacher M, Ruoslahti E. Cell attachment of repli- addition for the laboratory cas of SDS polyacrylamide gels reveals two adhesive to the methods http://jcp.bmj.com/ plasma proteins. J Cell Biol 1989;95:20-3. assessment of such patients in clinical prac- 24 Akama T, Yamada KM, Seno N, et al. Immunological tice. To understand the importance of characterization of human vitronectin and its binding to glycosaniinoglycans. J Biochem 1986;100: 1343-51. reduced serum concentrations of vitronectin 25 Conlan MG, Tomasini BR, Schultz RL, Mosher DF. in disease, however, and its possible implica- Plasma vitronectin polymorphism in normal subjects and patients with disseminated intravascular coagula- tion in pathogenic processes, a larger study is tion. Blood 1988;72:185-90. required. 26 Sun WH, Mosher DF. Polymorphism of vitronectin. Blood 1989;73:353-4.

27 Kubota K, Katayama S, Matsuda M, Hayashi M. Three on September 24, 2021 by guest. Protected copyright. types of vitronectin in human blood. Blood 1989;73: 353-4. 28 Preissner KT, Jenne D. Structure of vitronectin and its 1 Holmes R. Preparation from human serum of an alpha- biological role in haemostasis. Thromb Haemostas one protein which induces the immediate growth of 1991;66: 123-32. unadapted cells in vitro. J Cell Biol 1967;32:297-308. 29 Jenne D, Hugo F, Bhakdi S. Monoclonal antibodies to 2 Jenne D, Stanley KK. Molecular cloning of S-protein, human plasma Protein X alias complement S protein. a link between complement, coagulation and cell- Biosci Rep 1985;5:343-52. substrate adhesion. EMBO J7 1985;4:3153-7. 30 Brown C, Steen KS, Falk RJ, Woodley DT, O'Keefe, EJ. 3 Steen K. Epibolin: a protein of human plasma that sup- Vitronectin: effects on keratinocyte motility and inhibi- ports epithelial cell movement. Proc Nail Acad Sci USA tion of collagen induced motility. J Invest Dermatol 1981;78:6907-1 1. 199 1;96:724-8. 4 Shaffer MC, Foley TP, Barnes W. Quantitation of spread- 31 Dahlback K, Lofberg H, Alumets J, Dahlback B. ing factor in human biological fluids. J Lab Clin Med Immunohistochemical demonstration of age related 1984;103:783-91. depositions of vitronectin (S-protein of complement) 5 Hayman EG, Pierschbacher MD, Ohgren Y, Ruoslahti E. and terminal complement complex on dermal elastic Serum spreading factor (vitronectin) is present at the fibres. _lInvest Dermatol 1989;92:727-33. cell surface and in tissues. Proc Nail Acad Sci USA 32 Inuzuka S, Ueno T, Torimura T, et al. Vitronectin in liver 1983;80:4003-7. disorders: biochemical and immunohistochemical stud- 6 Podack ER, Muller-Eberhard J. Isolation of human S- ies. Hepatology 1992;15:629-36.