nephelometry as a comparison method for ranking the kits with regard to precision and accuracy. We also compared antibody and fibronectin cross reactivity.
CLIN.
CHEM.
33/1,
126-129
Four Immunoassay DanIel
B. Brubaker,
Kenneth
(1987)
Methods
and Standards
Compared
kits to meawith readily kits (Cooper
Biomedical and Boehringer-Mannheim) an immunoturbidimetric method is used. In a third kit (Biomedical Technologies, Inc.) an enzyme immunoassay method is used. To evaluate these commercial kits for fibronectin assay, we selected nephelometry as a comparison method for ranking the kits with regard to precision and accuracy. We also compared antibody and fibronectin cross reactivity. The antibodies from various manufacturers appear similar, but the fibronectin standards from different sources showed significant variation. Rate nephelometry and the BoehringerMannheim kit had the best within-run precision (CVs of 0.38% and 5.5% respectively). Between-run precision for nephelometry was excellent (CV = 1.9%) and somewhat high for the Boehringer-Mannheim kit (CV = 15.4%). This study demonstrates a need for (a) further standardization of antigen (fibronectin) and antibody in commercial kits and (b) the development of suitable stable quality-control material.
nephelometiy
Keyphrases: “kit” methods immunoturbidimetiy .
.
inter-kit results rate enzyme immunoassay
Assay of fibronectin in plasma and other body fluids is useful in evaluating the status of the mononuclear phagocytic system (1-5). Fibronectin, a dimeric alpha2-migrating surface-binding glycoprotein with a molecular mass of 440 000 Da (6), is decreased in plasma in patients after trauma (7), major surgery (8, 9), starvation (10), burns (11,
12), or sepsis (13), and in severely ill patients with evidence of disseminated intravascular coagulation (14). Reduced concentrations in plasma may be clinically significant because phagocytic dysfunction causes impaired circulatory clearance of particulate debris (15, 16). Currently, six immunoassay techniques are used for fibronectin assays: the electroimmunoassay (Laureh) method (17, 18), immunodiffusion (19), nephelometry (20, 21), enzyme-linked immunoassay (22), radioimmunoassay (19), and immunoturbidimetry (23). In the past, electroimmunoassay of fibronectin has been preferred; however, this method is relatively time consuming and expensive. Recently, several manufacturers have developed commercial kits for use in measuring fibronectin rapidly and inexpensively with readily available laboratory equipment. The reagent, antibodies, and fibronectin standards are critical for intercomparison of experimental and clinical results. Accordingly, we have evaluated four methods, including three commercially available kits, with different anti-fibronectin antibodies and fibronectin standards, assessing the accuracy, precision, and cross reactivity of antibodies and standards. The commercial kits included in this study are currently University
Pathology, Hospitals,
of Oklahoma Health Sciences Center, Department of College of Medicine, and State of Oklahoma Teaching P.O. Box 26307, Oklahoma City, OK 73126 (direct
correspondence to the latter address). Received July 7, 1986; accepted October
126
CLINICAL CHEMISTRY,
Fibronectin
E. BlIck, and MaxIne Romlne
Several companies have developed commercial sure plasma fibronectin rapidly and inexpensively available laboratory equipment. In two of these
Additional
for Measuring
Vol. 33, No.
14, 1986. 1, 1987
available in essentially paper, and our findings
these
kits
the may
or to new kits that
same
format
described in this to modifications of may become available in the
be applied
future.
Materials
and
Methods
Electroimmunoassay. Warmed antibodies to human fibronectin (Cooper Biomedical Inc., Malvern, PA 19355), 50 tg in 25 mL of a 10 g/L solution of agarose in a barbital buffer system, was poured into a 110 x 125 mm gel form. After the gel had hardened, we punched wells, 3 mm in diameter, in the aga.rose, 1.0cm apart. A 10-giL aliquot of each fraction to be tested for fibronectin or 10-ML standards of human fibronectin (Collaborative Research, Inc., Waltham, MA 02154) were added to the individual wells. The plate was electrophoresed at 80 V for 22 h at 15 #{176}C in an electrophoresis system with a power supply (Model 1420A) and an electrophoresis cell (Model 1405) from Bio-Rad, Richmond, CA 94804, and a temperature control (Poly Temp) from Polyscience, Warrington, PA 18976. The plates were blotted and rinsed with 0.15 mollL NaCl for 10 mm, then with de-ionized water for 10 mm. The gels were dried in an oven, then immersed in stain for 2 to 3 mm. This stain solution consisted of 5 g of Coomassie Brilliant
Blue R-250, 450 mL of 95% ethanol, 100 mL of acetic acid, and 450 mL of de-ionized water. The plates were immersed in three changes of destaining solution (the same mixture except Coomassie Brilliant Blue not included), 10 ruin each time, then allowed to dry. We used a caliper with 1.0-mm divisions to measure the rocket heights from the center of the well to the top of the rocket. Rocket heights for the standards were plotted and sample concentrations were calculated by use of this standard curve. Turbidimetry. We evaluated two immunoturbidimetric kits: “Plasma Fibronectin Assay System” (Cooper Biomedical) and the fibronectin kit of Boehringer-Mannheim Biochemistry Division, Indianapolis, IN 46250. The suppliers’ instructions were strictly followed. We prepared five fibronectin standards from each kit. Anti-human fibronectin antiserum, raised in goats (Cooper Biomedical) or sheep (Boehringer-Mannheim), was resuspended in buffer or distilled water, depending on the suppliers’ recommendation, then added to the standards, blank, and test samples. Kinetic turbidimetric measurement of the antigen-antibody reaction was recorded according to the principle of the fixedtime method for each kit; we used a Cary 219 double-beam spectrophotometer (Varian Associates, Palo Alto, CA 94303), set at 340 nm. Fibronectin concentrations in the standards were plotted as the observed absorbance difference; sample concentrations were determined from this calibration curve. Enzyme immunoassay (EIA). The EIA kit for fibronectin (Biomedical Technologies, Inc., Cambridge, MA 02141) involved a double-antibody technique with an alkaline phosphatase (EC 3.1.3.1)-labeled human fibronectin tracer. Detailed instructions supplied with the kit were strictly followed. Rabbit anti-fibronectin antiserum was added to
samples, standards, and controls. Alkaline phosphataselabeled human fibronectin was then added, followed by a precipitating antiserum (goat anti-rabbit gamma-globulin). All samples and standards were centrifuged (2000 X g, 4#{176}C, 30 mm). After decanting the supernates, we mixed the precipitates with chilled enzyme substrate (p-nitrophenyl phosphate) and incubated at 37 #{176}C for 30 mm, after which stopping reagent (KOH solution) was added. Absorbance was
measured
at 403 nm;
net
absorbance
was
determined
subtracting nonspecific binding values for all samples, plotting on semi-log paper vs log concentration.
by
then
Nephelometry. Our nephelometry methods are similar to those of Gressner and Wallraff (20). We used an ICS Analyzer II nephelometer (Beckman Instruments, Fullerton, CA 92634) and undiluted rabbit anti-human fibronectin antibody from Cooper Biomedical. Fibronectin standards were prepared in the following dilutions; six-, 10-, 15-, 20-,
30-, and 50-fold.
All reagents
lometer were identical were diluted 30-fold. computer programming
for
and materials each
experiment.
for the nepheAll
samples
Mode” feature with the sequence was followed in accordance with the manufacturer’s instructions for all measurements with the ICS. Pooled sera were included in each set of determinations as controls. To distinguish whether a peak rate was occurring in antigen excess in the Manual Mode procedure, we tested the unknown sample diluted 30-fold and 50-fold. If the 50fold dilution sample provided a greater rate response than the 30-fold dilution, which should have provided proportional results, the latter sample was judged to be in “antigen excess,” and the value obtained for it was considered invalid. The fibronectin concentration of each test specimen was determined from a linear standard curve of the peak rate values vs the corresponding antigen concentrations. Cross reactivity. We evaluated four anti-fibronectin antibodies from various sources for cross reactivity and standardization. Rabbit anti-human fibronectin used in the nephelometer method was obtained from Cooper Biomedical, whereas the Cooper fibronectin kit contained goat antihuman fibronectin. The Boehringer-Mannheim kit contains sheep anti-human fibronectin; the Biomedical Technologies kit contains rabbit anti-human fibronectin. Twenty vials of cryoprecipitate were thawed, aliquoted into 1-mL samples, and refrozen at -80 #{176}C. Fibronectin is concentrated in the cryoprecipitate, the cold-insoluble portion
of fresh-frozen
A “Manual
plasma
after
it has been
thawed
Results Our assessment of the cross reactivity of the anti-human fibronectin antibodies from different manufacturers indicated significant variation (Table 1). The turbidimetric methods in which we used the Boehringer-Mannheim kit with
fibronectin less variation.
same wavelength, the Cooper substantive differences between
anti-human fibronectin antibodies from various sources: 847.25 and 1817.8 pg/mL (Table 1), differences most probably ascribable to variation in the different antibody preparations and in the sensitivity of the methods used in the kits. Biomedical Technologies’ antibody could be used only with their kit, not in any other methods. When the amount of fibronectin is diluted within the reagent antibody range of Biomedical Technologies’ kit, the result is below the detection range of the spectrophotometer or nephelometer. Nephelometrically determined concentrations of fibronectin were consistently higher than by other methods, there being little variation with antibodies from different sources (Table 1). Variation in the amount of fibronectin analytically accounted for in the fibronectin standards was most significant between manufacturers (Table 2). None of the antibodies and methods yielded the exact or similar fibronectin standard value stated by the manufacturer. Nephelometry gave very high values for fibronectin standards when antisera from Cooper Biomedical or Boehringer Mannheim were used. Biomedical Technologies’ kit fibronectin standard was assayed at 150 jig/mL, much closer to the manufacturer’s stated value than was true for the other standards. Results
for the fibronectin standard from Biomedical Technologies were consistently closer to the stated value by all methods and with different antibodies, compared with results for the other suppliers’ standards. Also, assay of the Biomedical Technologies’ standards with the Boehringer-Mannheim kit yielded a value of 105 pg/rnL (stated value of 100 g/mL) the same as the stated value for the Boehringer-Mannheim low fibronectin standard. Within-run precision (Table 3) was excellent for the nephelometry method (CV = 0.38%) with Cooper Biomedical rabbit anti-human fibronectin and the Collaborative Research fibronectin standard. The Boehringer-Mannheim method had acceptable precision (CV = 5.5%). We did not assess within-run precision for cause that method permitted (agarose plate, antibody-antigen
immunoelectrophoresis beno variation in reagents reaction, buffer, etc.) and
equipment for all 20 samples. Nephelometry had the best between-run precision (CV = 1.9%), followed by immunoelectrophoresis (CV = 8.3%), then the Boehringer-Mannheim kit method (CV = 15.4%).
Table
1. Cross-Reactivity Human
antibodies from various sources However, in assays with the same
of Some Flbronectln Source
Cooper, rabbIt
Methodsa
BoehringerMannheim
13l9.5’
Cooper Biomedical Biomedical Technologies
three different assays.
showed
at the showed
at 1 to
6#{176}C. We assayed these samples for fibronectin, using each of the four antibodies with all five methods. We also assayed the fibronectin standards from each kit (Cooper Biomedical, Boehringer-Mannheim, and Biomedical Technologies) and the purified fibronectin we used as a standard, from Collaborative Research, by each of the five methods. Precision. To establish within-run precision, we assayed one sample of cryoprecipitate 20 times by each kit or method. To evaluate between-run precision, we quantified the fibronectin in the same 20 samples by all five methods in
anti-human
spectrophotometer Biomedical method
847.2 1113.75
Nephelometry a
Each
antibodies
kit
or method
to
of antibody
TechnologIes,
goat
BoehrlngerMannhelm, sheep
1322.5
1067.9
NDC
1685.6
1817.8
ND
NRd
NRd
1409.6
Cooper,
2713.3 2800.3 was used with its own #{176}Fibronectin concn,
BiomedIcal rabbit
2715.3 fibronectin
standards
ND and the
mean of results for 20 of the equipment. dNR, all results judged inaccurate because of a probable chemical interference from detergents and stabilizers in the anti-human fibronectin reagents. samples.
indicated.
AntIbodIes
g/mL,
CND, not detectable;below the sensitivity
CLINICAL CHEMISTRY,
Vol. 33, No. 1, 1987
127
Table
2. Fibronectln
Standards
a
Flbronectln Stated concn
Standard
Cooper Biomedical
440 655
895 Biomedical Technologies Collaborative
100
Cooper
BiomedIcal
Technologies
BIomedIcal
250 300 400 600
50 60
65 68
70
80
90
116
1500 50
150
150 68
205
750
600
125
224
1450 2000 2700 150
825 1080 1400 NR
475 750 950
570 780 825
105
NR
NR
450
300 500 105 215
Mannheim
BoehrlngerMannhelm
400 600 800 900 1700 600
200
Boehringer-
immunoelectrophoresIs
Nephelometryt”#{176}
100 150
Comparison
concn., mgIL
1000
-
-
160
875
400
Research NA, no results, concentration offibronectin standardis below the sensitivity of equipment. that kit’s anti-human fibronectin. bWlth nephelometry and lmmunoelectrophoresis.
rabbit anti-human
aThe fibronectin standards from each kit or method are assayed with antibody from Cooper Biomedical. eCollaborative Research Standard used in both
fibronectin
Table
3. Precision run
Between
Within run
cv,
Flbronectin. ,z/mL Kft or method
of Kits and Methods
Fibronectln,
Cv,
s.ag/mL
Mean
SD
25.4
779.2
52.8
6.8
164.4
15.4
643
35.2
5.5
348.9
20.7
1228
104.1
8.5
80.6 49.1
8.3 1.9
1895
Mean
SO
1409.5
357.3
1067.9
1685.5
Biomedical
Technologies
kit
(EIA)
Boehringer-Mannheim kit (turbidometric)
Cooper
Biomedical
kit
(turbidometric)
Immunoelectrophoresis
971.15 2516
Nephelometry n
=
studies suggested that the variation in accuracy between various methods and kit manufacturers was due to cross reactivity of the anti-human fibronectin antibodies from different manufacturers (24). Others have reported wide ranges of inter-species cross Results
of earlier
and
precision
antisera to fibronectin (19, variation was less significant methods (Table 1), the most variation being Cooper Biomedical kit. reactivity cross-reactive
for
greatest
source
of variation
between
However, for certain noted with the 25).
methods
was the
fibronectin standards. In general, our measured values for the fibronectin standards were not comparable with the values stated by the manufacturers. Only the Biomedical Technologies standard was close to its stated value (Table 2). Gressner and Wallraff have suggested that variations in the standards supplied by kit suppliers reflect differences in antigen concentration rather than differences in methods (20, 23). Our comparison of the fibronectin standard supports
this
observation:
two
different
turbidimetric
kits,
used
with the same spectrophotometer, gave significantly different results for their fibronectin standards. Because each fibronectin standard is apparently optimized for the specific kit method, problems arise when fibronectin obtained from one manufacturer is used with another supplier’s antibody or procedure. 128
7.3
0.38
20 each
DIscussion
The
Not tested
CLINICAL CHEMISTRY,
Vol. 33, No. 1, 1987
The variance in fibronectin standards is probably a result of the diverse antigemc character in the major forms of fibronectin (3,26,27). Fibronectin exists in human plasma in multiple molecular forms, distinguishable by sodium dodecyl sulfate/polyacrylamide gel electrophoresis (26). The particular form and the procedure used for purification of fibronectin no doubt account for variations in fibronectin standards from different suppliers. Our findings emphasize the need for international standardization of reagents, especially fibronectin standards and a suitable control sera. Precision studies showed wide variation among methods. The precision of the turbidimetric method was highly dependent on the quality of reagents. Our data suggest that Boehringer-Mannheim currently has a better antibody than Cooper-Biomedical for measuring fibronectin turbidimetrically. Indeed, the precision of the Boehringer-Mannheim kit most closely corresponded to that of the nephelometric method, in agreement with the findings of Gressner and Wallraff (23). Nephelometry was extremely precise (CV = 0.38%), with little between-run variation (CV = 1.9%). In summary:
immunoturbidimetry
measures
fibronectin
in plasma simply, rapidly, and precisely, but is highly influenced by the quality of antibody and standards. Nephelometry, also a simple, rapid method for measuring fibronectin, has a wider range of application, because it is so much more sensitive; we also find it to be more precise. The major disadvantage of nephelometry is the initial equipment cost.
References
1. Carsons
S, Lavietes BB, Diamond H. Factors influencing the of fibronectin into synovial fluid cryoproteins. J Lab
incorporation Clin Med 1983;102:722-31. 2. Kuusela P, Seppala M, Brock DJ, Ruusuvarra L, Ruoslahti Amniotic fluid fibronectin in normal pregnancy and pregnancies with anencephalic fetus. Biomedicine 1978;29:296-9.
3. Masher DF. Fibronectin. Prog Hemostas Thromb 4. Stathakis NE, Fauntas A, Tsianos E. Plasma normal
subjects
and
in
various
disease
states.
1980;5:111-51. fibronectin in J Clin Pathol
1981;34:504-8. 5. Zardi L, Destree A, Babe E, Isliker H. Human milk fibronectin: identification of fibronectin fragments by transfer of milk proteins from polyaciylamide gels to nitrocellulose sheets. Fed Eur Biochem Soc 1982;143:105-8.
6. Mosesson MW, Umfleet RA. The cold-insoluble globulin of human plasma. I. Purification, primary characterization, and relationship to fibrinogen and other cold-insoluble components. J Biol Chem 1970;245:5728-36. 7. Saba TM, Jaffe EA. Plasma fibronectin (opsonic glycoprotein). Its synthesis by vascular endothelial cells and role in cardiopulmonary integrity after trauma as related to reticuloendothelial function. Am J Med 1980;68:577-94. 8. Aronsen KF, Ekelund G, Kindmark CO, Laurell CB. Sequential changes of plasma proteins after surgical trauma. Scand J Clin Lab Invest 1972;29(suppl):127-36. 9. Scovill WA, Saba TM, Kaplan JE, Bernard HR, Powers SR. Disturbances in circulatory and blunt trauma. J Surg
opsomc activity Res 197722:709-16.
in man after
operative
10. Scott RL, Sohmer PR, Macdonald MG. The effect of starvation and repletion on plasma fibronectin in man. J Am Med Assoc
1982;2482025-7. 11. Goldman AS, Rudloff HB, McNamee R, Loose U), Diluzio NE. Deficiency of plasma humoral recognition factor activity following burn injury. J Reticuloendothelial Soc 1974;15:192-8. 12. Lancer ME, Saba TM, Scovill WA. Opsonic glycoprotein (plasma fibronectin) levels after burn injury: relationship to extent of burn and development of sepsis. Ann Surg 1980;192:776-82.
13. Lancer
ME, Saba TM. Opsonic fibronectin sepsis: cause or effect? Ann Surg 1982;195:340-5.
deficiency
14. Masher DF, Williams EM. Fibronectin concentration creased in plasma of severely ill patients with disseminated vascular coagulation. J Lab Clin Med 1978;91:729-35.
and is dointra-
15. Saba TM, Blumenstock FA, fibronectin and opsomc deficiency Ann Surg 1984;199:87-96.
Dhiraj MS, et al. Reversal of in patients. A controlled study.
16. Kaplan JE, Saba TM. Low grade intravascular coagulation reticulo-endothelial function. Ann J Physiol 1978;234:H323-9.
and
17. Blumenstock F, Weber P, Saba TM, Laffin R. Electroimmunoassay of alpha-2-opsonic protein during reticuloendothelial blockade. Am J Physiol 1977;323:R80-6. 18. Laurell CB. Quantitative estimation of proteins by electrophoresis in agarose gel containing antibodies. Anal Biochem 1966;150:45-52. 19. Kuusila P, Ruoslahti E, Engvall E, Valieri interspecies cross reactions of fibroblast surface tin). Immunochemistry 1976;13:639-42.
A. Immunological antigen
(fibronec-
20. Gressner AM, Wallraff P. Immunoturbidimetric measurement of cold insoluble globulin with reference to a laser nephelometric assay. J Clin Chem Clin Biochem 1981;19:1213-5.
21. Post G, Meyei-ing M, Voss B, Karges HE, Sieber A. Rapid determination of fibronectin by laser nephelometry. Fibronectin concentrations in plasma in human diseases, I. J Clin Chem Clin Biochem 1980;18:893-5. 22. Engvall surface
protein,
E, Ruoslahti fibronectin
E. Binding to collagen,
of soluble form of fibroblast hit J Cancer 1977;20:1-5.
23. Gressner AM, Wallraff P. Der Einsatz zur Bestimmung und rechnerunterstutzten bronectinkonzentration in verschiedenen Clin Chem Clin Biochem 1980;18:797-805.
der Lasernephelometrie Auswertung der FiKorperflussigkeiten. J
24. Brubaker DB, Romine M. Comparative precision analysis between four immunoassay methods determining fibronectin concentrations. 18th Cong hit Soc Blood Transfusions Abst 1984,P1220. 25. Ruoslahti E, Engvall properties of fibronectin.
E. Immunochemical and collagen-binding Ann NY Aced Sci 1978;312:178-91.
26. Amrani DL, Homandberg GA, Tooney NN, Wolfenstein-Todel C, Mosesson MW. Separation and analysis of the major forms of plasma flbronectin. Biochim Biophys Acta 1981;748:308-20. 27. Kavinsky CJ, Clark WA, Garber BB. linmunochemical sis of fibronectin using monoclonal antibodies. Biochim Acta
analyBiophys
1981;705:330-40.
CLINICAL CHEMISTRY,
Vol. 33, No. 1, 1987
129
