- Email: [email protected]
A two-site immunofluorometric assay for human placental lactogen
Clinica Chimicu Acta, Elsevier/North-Holland
CCA
I 14 (1981) I - 9 Biomedical
Press
1792
A two-site immunofluorometric assay for human placental lactogen L. Viinikka
*, J. Landon
and M. Pourfarzaneh
Department of Chemical Pathology, Si. Bartholomew’s Hospital, London EClA 7HL (U.K.) (Received
November
3rd, 1980)
suulmmy
A two-site immunofluorometric assay for human placental lactogen (HPL) in serum has been developed. Samples and standards are incubated for 10 min with an excess of sheep anti-HPL serum covalently coupled to particles of magnetisable cellulose. After sedimenting the particles and adsorbed hormone on a magnet the supemates are aspirated (or decanted) to waste. An excess of purified sheep anti-HPL immunoglobulin, labelled with fluorescein, is added and, after a further 20 min, the fluorescence remaining in the supernates, after sedimenting the particles, is inversely related to the initial concentration of HPL. Results correlate closely with those of an established radioimmunoassay (r = 0.92), within- and between-assay coefficients of variation are less than 10% and, employing a 200 ~1 sample volume, the assay extends from a minimum detection limit of 0.02 mg/l throughout the entire range of values encountered in pregnancy to more than 10 mg/l.
Introduction
Measurement of serum levels of human placental lactogen is widely performed to assess placental function during the third trimester [l-4] and as a guide in predicting the outcome of threatened abortion during early pregnancy [5]. Radioimmunoassays have been employed most commonly for this purpose [6-81 and non-isotopic immunoassays using HPL labelled with an enzyme [9] or a fluorophore [lo] have also been described. All immunoassays based on the use of a labelled antigen as tracer cover only a relatively narrow range of values. Thus it is necessary to set-up two, or even three standard curves [ 111, each with different reagent concentrations, to span the serum levels of HPL that are of clinical relevance between the fifth and final weeks of pregnancy-since these range from about 0.05 to 10 mg/l. Among their many advantages, immunoassays employing labelled antibodies cover a much wider range
WI. A two-site immunofluorometric assay (IFMA) for HPL has been developed using magnetisable particles for separation and specific anti-HPL immunoglobulins labelled
* To whom correspondence
0009-8981/8
should be addressed.
l/OOOO-0000/$02.50
0 Elsevier/North-Holland
Biomedical
Press
2 with fluorescein isothiocyanate (FITC) as tracer. The entire procedure, including fluorometry, is performed in disposable test tubes. Both early and late pregnancy samples can be assayed in a single run, centrifugation is avoided and use of a non-isotopic label offers several advantages.
Materials and methods HPL was from ICN Pharmaceuticals (Plainview, NY, U.S.A.3; FITC isomer I and cyanogen bromide (CNBr) from Sigma (London, U.K.); CNBr-activated Sepharose 4B and Sephadex G-25 fine grade from Pharmacia Fine Chemicals (London, U.K.); potassium isothiocyanate (KSCN) and sodium sulphate (Na,SO,) from BDH (Poole, Dorset, U.K.); disposable polystyrene test tubes (75 mm X 11 mm) from Walter Sarstedt (Leicester, U.K.) and magnetisable cellulose/ferric oxide particles and sheep anti-HPL serum from Technia Diagnostics (London, U.K.). The multipolar ferrite magnet was obtained from Magnet Applications (London, U.K.). Fluorimetry A Perkin-Elmer scribed previously
Model [ 131.
Buffers Bicarbonate buffer 7.5) were used.
1000 ratio-recording
(50 mmol/l,
filter
fluorimeter
pH 9.0) and phosphate
buffer
Preparation of unti-HPL mugnetisable cellulose particles Untreated sheep anti-HPL serum was coupled to magnetisable in the ratio 1 ml to 1 g, as described previously [14]. Preparation of HPL-immunoadsorbent HPL (5 mg) was coupled to CNBr-activated according to the manufacturer’s instructions.
Sepharose
was used as de-
(50 mmol/l,
cellulose
pH
particles,
4B solid phase (500 mg)
Preparution and purification of FITC-labelled anti-HPL immunoglobulins The immunoglobulin fraction from 1 ml of sheep anti-HPL serum was precipitated by addition of 180 mg of anhydrous Na,SO, with constant mixing for 30 min at room temperature, followed by centrifugation for 10 min at 2000 X g. The supernate was discarded, the precipitate washed twice with 2 ml of aqueous Na,SO, (180 g/l) and then dissolved in 1 ml of bicarbonate buffer. The protein content was determined by measurement of the optical density at 280 nm assuming an extinction coefficient of E$& = 15 and a molecular mass for sheep immunoglobulin G (IgG) of 145000 [15]. The sheep immunoglobulins and FITC were reacted overnight at room temperature in bicarbonate buffer at a molar ratio of 1 : 10 and then applied to a column (1.2 cm X 20 cm) of Sephadex G-25 and eluted with the bicarbonate buffer, collecting the entire labelled protein peak (5 ml) identified by its colour. This was then incubated at 4°C overnight with the HPL-immunoadsorbent (500 mg in 4 ml of phosphate buffer) while mixing on a rotator. The product was poured into a small column (0.7 cm X 10 cm), the unbound proteins washed to waste with 300 ml of the phosphate
3 buffer and the FITC-labelled specific anti-HPL immunoglobulins (FITC-anti-HPL) then eluted with 5 ml KSCN (3 mol/l). Finally the purified FITC-anti-HPL was separated from KSCN on a Sephadex G-25 column (as described above, but using phosphate buffer for elution) and stored in aliquots at -20°C with sodium azide (1 g/l) as preservative. Immunofluorometric assay procedure
Samples and standards (25 ~1 or 200 ~1) were incubated with excess anti-HPL magnetisable particles (2 mg in 100 ~1 of phosphate buffer) for 10 min and a further 1 ml of phosphate buffer added. The magnetisable particles, to which had been immunoadsorbed all the antigen, were sedimented on a magnet and the supernate, containing endogenous fluorophores and other interfering factors, aspirated to waste. An excess of the specific FITC-anti-HPL immunoglobulins (100 ~1 of the working dilution, equivalent to a 1 : 300 dilution of the original antiserum) was added and incubated for 20 min with one mix at 10 min. Finally, the fluorescence of the unbound FITC-anti-HPL was determined in the supernate after addition of 1.1 ml phosphate buffer and application of the magnet. Radioimmunoassay procedure
This was performed as described previously separate the antibody bound and free fractions.
[6] with ethanol being used to
Results Preparation of FITC-labelled anti-HPL immunoglobulins
In initial studies FITC and sheep immunoglobulin G were reacted in molar ratio’s of between 1 : 1 and 20 : 1. More than 90% incorporation of the fluorophore was obtained throughout but the fluorescent signal of the FITC labelled preparations did not show a linear relationship with the degree of labelling (Fig. 1). On the basis of these studies a molar ratio of 10: 1 was employed to produce FITC-anti-HPL for assay purposes and the immunoreactivity of the labelled product was shown not to change during six months storage in phosphate buffer at -20°C. The amount of FITC-anti-HPL obtained from one ml of the original sheep antiserum after elution from the immunoadsorbent with KSCN was sufficient for more than 3000 determinations. Additional elutions did not recover significant further amounts of labelled antibody. Kinetic studies
The time taken to attain equilibrium during the first incubation was determined by incubating 25 ~1 of male serum containing either 1 mg/l or 10 mg/l of HPL with 100 ~1 (2 mg) of sheep anti-HPL serum covalently linked to magnetisable particles for from 2 to 30 min. The second incubation period was kept constant. The results (Fig. 2a) indicate that all the HPL in the sera had been immunoadsorbed within 10 min. The time taken to reach equilibrium after addition of the FITC-anti-HPL immunoglobulins was also studied as above, but keeping the time for the first incubation constant at 10 rnin and varying the second incubation time from 2 to 30 min. Equilibrium was attained within 20 min (Fig. 2b).
OLl 2
6 Rat10 of FITC
10
~_-.14
18
to lmmunoglobulinG
Fig. 1. Fluorescence, expressed in arbitrary units, of similar concentrations of sheep immunoglobulin ti each labelled with a different molar ratio of FITC ranging from I : 20. Above a molar ratio of about I : 4 there was not a linear relationship between fluorescence and the number of molecules of FITC per molecule of immunoglobuiin G due to concentration quenching.
standard curves and assay characteristics
Representative standard curves employing 25 ~1 and 200 ~1 of HPL standards in normal male serum are shown in Fig. 3. That using a 25 ~1 volume covered the range of values of clinical interest during the third trimester of pregnancy. Use of 200 ~1
Fig. 2. Kinetics of the two reactions involved in the two-site ~mrnuno~uorom~t~c assay using 25 gf of serum standards containing I mg/l (O------O) or IO mg/l (O0) of HPL. (a) The first reaction in which HPL is being bound to sheep anti-HPL serum covalently coupled to magnetisable particles and (b) the second reaction after the addition of FITC-labelled specific anti-HPL immunoglobulins. As a result times of 10 min and 20 min respectively were chosen for the assay.
5
Fig. 3. Representative based standards.
standard
curves obtained
using 25 yl (O------a)
or 21X)pi (I -0)
of serun1
provided a standard curve which spanned all levels of clinical relevance encountered throughout pregnancy. Sensitivity was determined by measuring 20 replicates of the zero concentration standard, as described by Rodbard [ 161. The minimum detectable concentration of HPL in serum, at the 95% confidence limits, was 0.1 mg/l with a 25 ~1 sample volume. This was reduced to 0.02 mg/l when a 200 ~1 sample volume was used. The within- and between-assay coefficients of variation were determined by measuring two pools of pregnancy sera 11 times within the same assay and in 11 different assays. At mean values of 1.6 mg/l and 8.2 mg/l the within-assay values were 9.5% and 5.98, respectively, and the between-assay C.V. 9.7% and 8.7%. The
TABLE
1
PRECISION
OF END-POINT
DETECTION
OF FITC-ANTI-HPL
I tube measured x20
IMMUNOGLOBULINS
20 tubes measured once
Pnlystyrene (Sarstedt
assay tubes 55.478)
71.4k0.32
(0.45%) *
72.4&0.71
(0.988)
Polystyrene (Sarstqlt
cuvettes 67.754)
95.9 kO.27
(0.29%)
96.8iO.33
(0.34%)
84.450.28
(0.33%)
87.4iO.81
(0.93%)
Glass cuvettes (Hellma 6080F)
* Fluorescence reading: mean ?I S.D. (C.V.). The differences in the signal obtained reflect differences in the diameter of the tube or cuvette. These were 5 mmX 5 mm, 10 mmX 10 mm. and IO mmX4 mm for the tubes, the polystyrene cuvettes. and the glass cuvettes, respectively.
6 12r
2 HPL Levels
4 (mgil)
6
8
10
12
by Rodloimmunoassay
Fig. 4. Correlation between the levels of HPL in pregnant women’s samples as obtained by the two-site IFMA ( ,V axis) and a conventional radioimmunoassay (x axis). II = 62, J = I .09x - I .20, r= 0.92.
contribution to imprecision by fluorometry and the measurement of fluorescence within the polystyrene assay tubes was assessed. The amount of labelled specific immunoglobulins employed in the assay was pipetted, with phosphate buffer, into 20 assay tubes and the fluorescence determined. Twenty readings of one tube gave a C.V. of 0.45% and this was increased to only 0.98% for the 20 tubes, indicating the precision of end-point detection and a minimal contribution by the use of assay tubes for fluorometry. Fluorescence values obtained in polystyrene cuvettes and glass cuvettes were higher and precision somewhat better (Table I). Analytical recoveries based on serial dilutions of a high (16 mg/l) sample standard and from adding 2 or 5 mg/l of HPL to normal male sera ranged from 98.1 to 107.7%. Correlation of results with those by radioimmunoassay
Sera from 62 pregnant women were assayed by the IFMA (y) (using a 25 ~1 volume) and an established radioimmunoassay (x). The correlation coefficient was 0.92 and the results were related by the regression equation y = 1.09x - 1.20, calculated on the assumption that both methods had equal precision characteristics 1171. Discussion
The present two-site IFMA offers three benefits as compared with the conventional radioimmunoassays most commonly employed for the estimation of HPL levels in serum. First, the use of magnetisable particles avoids the need for centrifugation and greatly simplifies and speeds the separation and wash steps. Second, immunoassays employing labelled antibodies have several inherent advantages as compared with those based on the use of labelled antigen as the tracer. Finally, non-isotopic labels offer several advantages as compared with gamma-emitting radioisotopes, such as ‘25I especially for the labelling of immunoglobulins. Table II summarises some of the fundamental differences between immunoassays
7 TABLE II COMPARISON OF RADIOIMMUNOASSAY
AND TWO-SITE IMMUNORADIOMETRIC
Radioimmunoassay
Two-site (sandwich) immunoradiometric
Employs tracer-labelled antigen
Employs tracer-labelled antibody
Separate bound and free antigens
Separate bound and free antibodies
Counts in bound fraction inversely related to initial concentration of antigen
Counts in bound fraction dire& related to initial concentration of antigen
Limited reagent procedure, therefore: relatively slow
Excess reagent procedure, therefore: relatively rapid
precision of pipetting sample (or standard) labelled antigen and antiserum is critical
precision of pipetting is critical only for the specimen (and standard)
ultimate sensitivity directly related to the avidity of predominant antibodies
sensitivity less dependent upon antibody avidity and is set by the specific activity of the labelled antibodies
ASSAY
assay
Specificity directly related to that of a single antiserum
Enhanced specificity related to that of two different antibody populations
Covers relatively small range of analyte concentrations
Covers a wide range of analyte concentrations
Problems in labelling some antigens and short shelf life of labelled product
No problems in labelling immunoglobulins
Employs antisera at considerable dilutions
Requires isolated specific antibodies in relatively large amounts
employing labelled antigen and those employing labelled antibodies. Some of the advantages of the latter are seen in the present study. These include the much wider range of analyte concentrations covered in a single assay; the stability of the labelled reactant; sensitivity and the relative speed associated with employing a reagent excess. In the latter context, the total incubation time was 30 min for the present IFMA as compared with overnight for a conventional radioimmunoassay optimised to attain equivalent sensitivity [6]. Two disadvantages of immunoradiometric assays as compared to radioimmunoassays are that relatively larger amounts of antisera are required and the production of stable, specific ‘251-labelled antibodies is difficult technically. The former is not a problem when large animals, such as sheep, are immunised or the antigen, such as HPL, is a potent immunogen and high titres of antibodies are produced. However, radioiodination of specific antibodies requires considerable technical expertise and the labelled product is relatively unstable with a useful shelf-life of less than ten weeks [ 181.This is due to the incorporation of radioactive molecules rather than the iodination procedure since antibodies labelled with 12’1 are stable and retain immunoreactivity at high incorporation levels (G.H. Addison, personal communication). Finally, antibodies are usually radioiodinated when bound to an immunoadsorbent to prevent incorporation of label into the binding sites. This may be wasteful
8
of the antigen used as immunoadsorbent if it is also radioiodinated and cannot be reused. Added to such advantages of fluorophores as the speed and precision of end-point detection, it is simple to label antibodies with FITC. The slightly alkaline reaction conditions do not impair immunoreactivity and incorporation is virtually complete at molar ratios in excess of 20 : 1-although there is not a linear increase in signal with increasing numbers of fluorophores because of concentration quenching by a dipole-dipole mechanism [ 191.Further, it is possible to label all1the immunoglobulins and then immunoadsorb out the 1W, or less, that are specific to the antigen to be assayed. This is impossible in the case of radioiodination because of the cost and health hazard that would be associated with the 100 mCi or more of radioisotope that would be required. Finally, the fluorophore-labelled specific antibodies have a virtually indefinite shelf-life. A conventional (as opposed to a two-site) IFMA has also been developed in which the sample or standard was incubated with FITC-labelled specific anti-HPL immunoglobulins, the unbound labelled antibodies removed by addition of an excess of HPL covalently linked to a solid phase support and measurement of the fluorescence of the bound fraction in the supernate. This, like a similar assay employing enzyme-labelled antibodies [20], proved suitable for use with buffer but not for clinical samples- because endogenous factors present in serum interfered with end-point detection. A two-site IFMA is preferred because of the potentially improved specificity, the ease of removing endogenous fluorophores and other interfering factors during the separation step and the use of antibodies rather than antigen on the immunoabsorbent. Acknowledgements
One of the authors (L.V.) is most grateful for a grant from the Royal Society and the Academy of Finland. References I Chard, T. (1974) The fetus at risk. Lancet ii, 880-883 2 Spellacy, W.N., Buki, W.C. and Birk. S.A. (1975) The effectiveness of human placental lactogen measurements as an adjunct in decreasing perinatal deaths. Am. J. Obstet. Gynecol. 121, 835-843 3 Chard, T. (1976) Assessment of fetoplacental function by biochemical determinations. J. Clin. Pathol. 29, Suppl. IO. 18-26 4 Letchworth, A.T.. Slattery. M. and Dennis, K.J. (1978) Clinical application of human-placentallactogen values in late pregnancy. Lancet i, 955-957 5 Niven. P.A.R., Landon, J. and Chard, T. (1972) Placental lactogen levels as guide to outcome of threatened abortion. Br. Med. J. 3. 799-801 6 Letchworth, A.T.. Boardman, R., Bristow, C.. Landon, J. and Chard, T. (1971) A rapid radioimmunoassay for human chorionic somatomammotrophin. J. Obstet. Gynaecol. Br. Cwlth. 79, 535-541 7 Gardner, J., Bailey, G. and Chard, T. (1974) Observations on the use of solid-phase-coupled antibodies in the radioimmunoassay of human placental lactogen. Biochem. J. 137, 469-476 8 Ermshar. C.L. and Gusseck. D.J. (1978) Use of polyethylene glycol in radioimmunoassay of human placental lactogen. Clin. Chem. 24, l767- I769 9 Van Hell. H.. Brande, J.A.M. and Schuurs, A.H.W.M. (1979) Enzyme-immunoassay of human placental lactogen. Clin. Chim. Acta 91, 309-316 IO Chard, T. and Sykes, A. (1979) Fluoroimmunoassay for human choriomammotrophin. Clin. Chem. 25. 973-975
9 11 Radiochemical Centre (1980) HPL Immunoassay Kit. Amersham, Bucks, U.K. 12 Ekins, R.P. (1978) General principles of hormone assay. In: Hormone Assays and their Clinical Application (Loraine, J.A. and Bell, E.T., eds), Churchill-Livingston 13 Kamel, R.S., Landon, J. and Smith, D.S. (1980) Magnetisable solid-phase fluoroimmunoassay of phenytoin in disposable test tubes. Clin. Chem. 26, 1281- 1284 14 Pourfarzaneh, M., White, G.W., Landon, J. and Smith, D.S. (1980) Cortisol directly determined in serum by fluoroimmunoassay with magnetisable solid-phase. Clin. Chem. 26,730- 733 15 Ullman, E.F., Schwarzberg, M. and Rubenstein, K.E. (1976) Fluorescent excitation transfer immunoassay. A general method for determination of antigens. J. Biol. Chem. 251, 4172-4178 16 Rodbard, D. (1978) Statistical estimation of the minimal detectable concentration (“Sensitivity”) for radioligand assays. Anal. B&hem. 90, I- 12 17 Combleet, P.J. and Goohman, N. (1979) Incorrect least-square regression coefficients in method analysis. Clin. Chem. 25, 432-438 18 Woodhead, J.S., Addison, G.M. and Hales, C.N. (1974)The immunoradiometric assay and related techniques. Br. Med. Bull. 30, 38-49 19 Nargessi, R.D., Landon, J., Pourfarzaneh, M. and Smith, D.S. (1978) Solid-phase fluoroimmunoassay of human albumin in biological fluids. Clin. Chim. Acta 89, 455-460 20 Barbour, H.M. (1976) Development of an enzyme immunoassay for human placental lactogen using labelled antibodies. J. Immunol. Methods 1I, 15-23
CCA
I 14 (1981) I - 9 Biomedical
Press
1792
A two-site immunofluorometric assay for human placental lactogen L. Viinikka
*, J. Landon
and M. Pourfarzaneh
Department of Chemical Pathology, Si. Bartholomew’s Hospital, London EClA 7HL (U.K.) (Received
November
3rd, 1980)
suulmmy
A two-site immunofluorometric assay for human placental lactogen (HPL) in serum has been developed. Samples and standards are incubated for 10 min with an excess of sheep anti-HPL serum covalently coupled to particles of magnetisable cellulose. After sedimenting the particles and adsorbed hormone on a magnet the supemates are aspirated (or decanted) to waste. An excess of purified sheep anti-HPL immunoglobulin, labelled with fluorescein, is added and, after a further 20 min, the fluorescence remaining in the supernates, after sedimenting the particles, is inversely related to the initial concentration of HPL. Results correlate closely with those of an established radioimmunoassay (r = 0.92), within- and between-assay coefficients of variation are less than 10% and, employing a 200 ~1 sample volume, the assay extends from a minimum detection limit of 0.02 mg/l throughout the entire range of values encountered in pregnancy to more than 10 mg/l.
Introduction
Measurement of serum levels of human placental lactogen is widely performed to assess placental function during the third trimester [l-4] and as a guide in predicting the outcome of threatened abortion during early pregnancy [5]. Radioimmunoassays have been employed most commonly for this purpose [6-81 and non-isotopic immunoassays using HPL labelled with an enzyme [9] or a fluorophore [lo] have also been described. All immunoassays based on the use of a labelled antigen as tracer cover only a relatively narrow range of values. Thus it is necessary to set-up two, or even three standard curves [ 111, each with different reagent concentrations, to span the serum levels of HPL that are of clinical relevance between the fifth and final weeks of pregnancy-since these range from about 0.05 to 10 mg/l. Among their many advantages, immunoassays employing labelled antibodies cover a much wider range
WI. A two-site immunofluorometric assay (IFMA) for HPL has been developed using magnetisable particles for separation and specific anti-HPL immunoglobulins labelled
* To whom correspondence
0009-8981/8
should be addressed.
l/OOOO-0000/$02.50
0 Elsevier/North-Holland
Biomedical
Press
2 with fluorescein isothiocyanate (FITC) as tracer. The entire procedure, including fluorometry, is performed in disposable test tubes. Both early and late pregnancy samples can be assayed in a single run, centrifugation is avoided and use of a non-isotopic label offers several advantages.
Materials and methods HPL was from ICN Pharmaceuticals (Plainview, NY, U.S.A.3; FITC isomer I and cyanogen bromide (CNBr) from Sigma (London, U.K.); CNBr-activated Sepharose 4B and Sephadex G-25 fine grade from Pharmacia Fine Chemicals (London, U.K.); potassium isothiocyanate (KSCN) and sodium sulphate (Na,SO,) from BDH (Poole, Dorset, U.K.); disposable polystyrene test tubes (75 mm X 11 mm) from Walter Sarstedt (Leicester, U.K.) and magnetisable cellulose/ferric oxide particles and sheep anti-HPL serum from Technia Diagnostics (London, U.K.). The multipolar ferrite magnet was obtained from Magnet Applications (London, U.K.). Fluorimetry A Perkin-Elmer scribed previously
Model [ 131.
Buffers Bicarbonate buffer 7.5) were used.
1000 ratio-recording
(50 mmol/l,
filter
fluorimeter
pH 9.0) and phosphate
buffer
Preparation of unti-HPL mugnetisable cellulose particles Untreated sheep anti-HPL serum was coupled to magnetisable in the ratio 1 ml to 1 g, as described previously [14]. Preparation of HPL-immunoadsorbent HPL (5 mg) was coupled to CNBr-activated according to the manufacturer’s instructions.
Sepharose
was used as de-
(50 mmol/l,
cellulose
pH
particles,
4B solid phase (500 mg)
Preparution and purification of FITC-labelled anti-HPL immunoglobulins The immunoglobulin fraction from 1 ml of sheep anti-HPL serum was precipitated by addition of 180 mg of anhydrous Na,SO, with constant mixing for 30 min at room temperature, followed by centrifugation for 10 min at 2000 X g. The supernate was discarded, the precipitate washed twice with 2 ml of aqueous Na,SO, (180 g/l) and then dissolved in 1 ml of bicarbonate buffer. The protein content was determined by measurement of the optical density at 280 nm assuming an extinction coefficient of E$& = 15 and a molecular mass for sheep immunoglobulin G (IgG) of 145000 [15]. The sheep immunoglobulins and FITC were reacted overnight at room temperature in bicarbonate buffer at a molar ratio of 1 : 10 and then applied to a column (1.2 cm X 20 cm) of Sephadex G-25 and eluted with the bicarbonate buffer, collecting the entire labelled protein peak (5 ml) identified by its colour. This was then incubated at 4°C overnight with the HPL-immunoadsorbent (500 mg in 4 ml of phosphate buffer) while mixing on a rotator. The product was poured into a small column (0.7 cm X 10 cm), the unbound proteins washed to waste with 300 ml of the phosphate
3 buffer and the FITC-labelled specific anti-HPL immunoglobulins (FITC-anti-HPL) then eluted with 5 ml KSCN (3 mol/l). Finally the purified FITC-anti-HPL was separated from KSCN on a Sephadex G-25 column (as described above, but using phosphate buffer for elution) and stored in aliquots at -20°C with sodium azide (1 g/l) as preservative. Immunofluorometric assay procedure
Samples and standards (25 ~1 or 200 ~1) were incubated with excess anti-HPL magnetisable particles (2 mg in 100 ~1 of phosphate buffer) for 10 min and a further 1 ml of phosphate buffer added. The magnetisable particles, to which had been immunoadsorbed all the antigen, were sedimented on a magnet and the supernate, containing endogenous fluorophores and other interfering factors, aspirated to waste. An excess of the specific FITC-anti-HPL immunoglobulins (100 ~1 of the working dilution, equivalent to a 1 : 300 dilution of the original antiserum) was added and incubated for 20 min with one mix at 10 min. Finally, the fluorescence of the unbound FITC-anti-HPL was determined in the supernate after addition of 1.1 ml phosphate buffer and application of the magnet. Radioimmunoassay procedure
This was performed as described previously separate the antibody bound and free fractions.
[6] with ethanol being used to
Results Preparation of FITC-labelled anti-HPL immunoglobulins
In initial studies FITC and sheep immunoglobulin G were reacted in molar ratio’s of between 1 : 1 and 20 : 1. More than 90% incorporation of the fluorophore was obtained throughout but the fluorescent signal of the FITC labelled preparations did not show a linear relationship with the degree of labelling (Fig. 1). On the basis of these studies a molar ratio of 10: 1 was employed to produce FITC-anti-HPL for assay purposes and the immunoreactivity of the labelled product was shown not to change during six months storage in phosphate buffer at -20°C. The amount of FITC-anti-HPL obtained from one ml of the original sheep antiserum after elution from the immunoadsorbent with KSCN was sufficient for more than 3000 determinations. Additional elutions did not recover significant further amounts of labelled antibody. Kinetic studies
The time taken to attain equilibrium during the first incubation was determined by incubating 25 ~1 of male serum containing either 1 mg/l or 10 mg/l of HPL with 100 ~1 (2 mg) of sheep anti-HPL serum covalently linked to magnetisable particles for from 2 to 30 min. The second incubation period was kept constant. The results (Fig. 2a) indicate that all the HPL in the sera had been immunoadsorbed within 10 min. The time taken to reach equilibrium after addition of the FITC-anti-HPL immunoglobulins was also studied as above, but keeping the time for the first incubation constant at 10 rnin and varying the second incubation time from 2 to 30 min. Equilibrium was attained within 20 min (Fig. 2b).
OLl 2
6 Rat10 of FITC
10
~_-.14
18
to lmmunoglobulinG
Fig. 1. Fluorescence, expressed in arbitrary units, of similar concentrations of sheep immunoglobulin ti each labelled with a different molar ratio of FITC ranging from I : 20. Above a molar ratio of about I : 4 there was not a linear relationship between fluorescence and the number of molecules of FITC per molecule of immunoglobuiin G due to concentration quenching.
standard curves and assay characteristics
Representative standard curves employing 25 ~1 and 200 ~1 of HPL standards in normal male serum are shown in Fig. 3. That using a 25 ~1 volume covered the range of values of clinical interest during the third trimester of pregnancy. Use of 200 ~1
Fig. 2. Kinetics of the two reactions involved in the two-site ~mrnuno~uorom~t~c assay using 25 gf of serum standards containing I mg/l (O------O) or IO mg/l (O0) of HPL. (a) The first reaction in which HPL is being bound to sheep anti-HPL serum covalently coupled to magnetisable particles and (b) the second reaction after the addition of FITC-labelled specific anti-HPL immunoglobulins. As a result times of 10 min and 20 min respectively were chosen for the assay.
5
Fig. 3. Representative based standards.
standard
curves obtained
using 25 yl (O------a)
or 21X)pi (I -0)
of serun1
provided a standard curve which spanned all levels of clinical relevance encountered throughout pregnancy. Sensitivity was determined by measuring 20 replicates of the zero concentration standard, as described by Rodbard [ 161. The minimum detectable concentration of HPL in serum, at the 95% confidence limits, was 0.1 mg/l with a 25 ~1 sample volume. This was reduced to 0.02 mg/l when a 200 ~1 sample volume was used. The within- and between-assay coefficients of variation were determined by measuring two pools of pregnancy sera 11 times within the same assay and in 11 different assays. At mean values of 1.6 mg/l and 8.2 mg/l the within-assay values were 9.5% and 5.98, respectively, and the between-assay C.V. 9.7% and 8.7%. The
TABLE
1
PRECISION
OF END-POINT
DETECTION
OF FITC-ANTI-HPL
I tube measured x20
IMMUNOGLOBULINS
20 tubes measured once
Pnlystyrene (Sarstedt
assay tubes 55.478)
71.4k0.32
(0.45%) *
72.4&0.71
(0.988)
Polystyrene (Sarstqlt
cuvettes 67.754)
95.9 kO.27
(0.29%)
96.8iO.33
(0.34%)
84.450.28
(0.33%)
87.4iO.81
(0.93%)
Glass cuvettes (Hellma 6080F)
* Fluorescence reading: mean ?I S.D. (C.V.). The differences in the signal obtained reflect differences in the diameter of the tube or cuvette. These were 5 mmX 5 mm, 10 mmX 10 mm. and IO mmX4 mm for the tubes, the polystyrene cuvettes. and the glass cuvettes, respectively.
6 12r
2 HPL Levels
4 (mgil)
6
8
10
12
by Rodloimmunoassay
Fig. 4. Correlation between the levels of HPL in pregnant women’s samples as obtained by the two-site IFMA ( ,V axis) and a conventional radioimmunoassay (x axis). II = 62, J = I .09x - I .20, r= 0.92.
contribution to imprecision by fluorometry and the measurement of fluorescence within the polystyrene assay tubes was assessed. The amount of labelled specific immunoglobulins employed in the assay was pipetted, with phosphate buffer, into 20 assay tubes and the fluorescence determined. Twenty readings of one tube gave a C.V. of 0.45% and this was increased to only 0.98% for the 20 tubes, indicating the precision of end-point detection and a minimal contribution by the use of assay tubes for fluorometry. Fluorescence values obtained in polystyrene cuvettes and glass cuvettes were higher and precision somewhat better (Table I). Analytical recoveries based on serial dilutions of a high (16 mg/l) sample standard and from adding 2 or 5 mg/l of HPL to normal male sera ranged from 98.1 to 107.7%. Correlation of results with those by radioimmunoassay
Sera from 62 pregnant women were assayed by the IFMA (y) (using a 25 ~1 volume) and an established radioimmunoassay (x). The correlation coefficient was 0.92 and the results were related by the regression equation y = 1.09x - 1.20, calculated on the assumption that both methods had equal precision characteristics 1171. Discussion
The present two-site IFMA offers three benefits as compared with the conventional radioimmunoassays most commonly employed for the estimation of HPL levels in serum. First, the use of magnetisable particles avoids the need for centrifugation and greatly simplifies and speeds the separation and wash steps. Second, immunoassays employing labelled antibodies have several inherent advantages as compared with those based on the use of labelled antigen as the tracer. Finally, non-isotopic labels offer several advantages as compared with gamma-emitting radioisotopes, such as ‘25I especially for the labelling of immunoglobulins. Table II summarises some of the fundamental differences between immunoassays
7 TABLE II COMPARISON OF RADIOIMMUNOASSAY
AND TWO-SITE IMMUNORADIOMETRIC
Radioimmunoassay
Two-site (sandwich) immunoradiometric
Employs tracer-labelled antigen
Employs tracer-labelled antibody
Separate bound and free antigens
Separate bound and free antibodies
Counts in bound fraction inversely related to initial concentration of antigen
Counts in bound fraction dire& related to initial concentration of antigen
Limited reagent procedure, therefore: relatively slow
Excess reagent procedure, therefore: relatively rapid
precision of pipetting sample (or standard) labelled antigen and antiserum is critical
precision of pipetting is critical only for the specimen (and standard)
ultimate sensitivity directly related to the avidity of predominant antibodies
sensitivity less dependent upon antibody avidity and is set by the specific activity of the labelled antibodies
ASSAY
assay
Specificity directly related to that of a single antiserum
Enhanced specificity related to that of two different antibody populations
Covers relatively small range of analyte concentrations
Covers a wide range of analyte concentrations
Problems in labelling some antigens and short shelf life of labelled product
No problems in labelling immunoglobulins
Employs antisera at considerable dilutions
Requires isolated specific antibodies in relatively large amounts
employing labelled antigen and those employing labelled antibodies. Some of the advantages of the latter are seen in the present study. These include the much wider range of analyte concentrations covered in a single assay; the stability of the labelled reactant; sensitivity and the relative speed associated with employing a reagent excess. In the latter context, the total incubation time was 30 min for the present IFMA as compared with overnight for a conventional radioimmunoassay optimised to attain equivalent sensitivity [6]. Two disadvantages of immunoradiometric assays as compared to radioimmunoassays are that relatively larger amounts of antisera are required and the production of stable, specific ‘251-labelled antibodies is difficult technically. The former is not a problem when large animals, such as sheep, are immunised or the antigen, such as HPL, is a potent immunogen and high titres of antibodies are produced. However, radioiodination of specific antibodies requires considerable technical expertise and the labelled product is relatively unstable with a useful shelf-life of less than ten weeks [ 181.This is due to the incorporation of radioactive molecules rather than the iodination procedure since antibodies labelled with 12’1 are stable and retain immunoreactivity at high incorporation levels (G.H. Addison, personal communication). Finally, antibodies are usually radioiodinated when bound to an immunoadsorbent to prevent incorporation of label into the binding sites. This may be wasteful
8
of the antigen used as immunoadsorbent if it is also radioiodinated and cannot be reused. Added to such advantages of fluorophores as the speed and precision of end-point detection, it is simple to label antibodies with FITC. The slightly alkaline reaction conditions do not impair immunoreactivity and incorporation is virtually complete at molar ratios in excess of 20 : 1-although there is not a linear increase in signal with increasing numbers of fluorophores because of concentration quenching by a dipole-dipole mechanism [ 191.Further, it is possible to label all1the immunoglobulins and then immunoadsorb out the 1W, or less, that are specific to the antigen to be assayed. This is impossible in the case of radioiodination because of the cost and health hazard that would be associated with the 100 mCi or more of radioisotope that would be required. Finally, the fluorophore-labelled specific antibodies have a virtually indefinite shelf-life. A conventional (as opposed to a two-site) IFMA has also been developed in which the sample or standard was incubated with FITC-labelled specific anti-HPL immunoglobulins, the unbound labelled antibodies removed by addition of an excess of HPL covalently linked to a solid phase support and measurement of the fluorescence of the bound fraction in the supernate. This, like a similar assay employing enzyme-labelled antibodies [20], proved suitable for use with buffer but not for clinical samples- because endogenous factors present in serum interfered with end-point detection. A two-site IFMA is preferred because of the potentially improved specificity, the ease of removing endogenous fluorophores and other interfering factors during the separation step and the use of antibodies rather than antigen on the immunoabsorbent. Acknowledgements
One of the authors (L.V.) is most grateful for a grant from the Royal Society and the Academy of Finland. References I Chard, T. (1974) The fetus at risk. Lancet ii, 880-883 2 Spellacy, W.N., Buki, W.C. and Birk. S.A. (1975) The effectiveness of human placental lactogen measurements as an adjunct in decreasing perinatal deaths. Am. J. Obstet. Gynecol. 121, 835-843 3 Chard, T. (1976) Assessment of fetoplacental function by biochemical determinations. J. Clin. Pathol. 29, Suppl. IO. 18-26 4 Letchworth, A.T.. Slattery. M. and Dennis, K.J. (1978) Clinical application of human-placentallactogen values in late pregnancy. Lancet i, 955-957 5 Niven. P.A.R., Landon, J. and Chard, T. (1972) Placental lactogen levels as guide to outcome of threatened abortion. Br. Med. J. 3. 799-801 6 Letchworth, A.T.. Boardman, R., Bristow, C.. Landon, J. and Chard, T. (1971) A rapid radioimmunoassay for human chorionic somatomammotrophin. J. Obstet. Gynaecol. Br. Cwlth. 79, 535-541 7 Gardner, J., Bailey, G. and Chard, T. (1974) Observations on the use of solid-phase-coupled antibodies in the radioimmunoassay of human placental lactogen. Biochem. J. 137, 469-476 8 Ermshar. C.L. and Gusseck. D.J. (1978) Use of polyethylene glycol in radioimmunoassay of human placental lactogen. Clin. Chem. 24, l767- I769 9 Van Hell. H.. Brande, J.A.M. and Schuurs, A.H.W.M. (1979) Enzyme-immunoassay of human placental lactogen. Clin. Chim. Acta 91, 309-316 IO Chard, T. and Sykes, A. (1979) Fluoroimmunoassay for human choriomammotrophin. Clin. Chem. 25. 973-975
9 11 Radiochemical Centre (1980) HPL Immunoassay Kit. Amersham, Bucks, U.K. 12 Ekins, R.P. (1978) General principles of hormone assay. In: Hormone Assays and their Clinical Application (Loraine, J.A. and Bell, E.T., eds), Churchill-Livingston 13 Kamel, R.S., Landon, J. and Smith, D.S. (1980) Magnetisable solid-phase fluoroimmunoassay of phenytoin in disposable test tubes. Clin. Chem. 26, 1281- 1284 14 Pourfarzaneh, M., White, G.W., Landon, J. and Smith, D.S. (1980) Cortisol directly determined in serum by fluoroimmunoassay with magnetisable solid-phase. Clin. Chem. 26,730- 733 15 Ullman, E.F., Schwarzberg, M. and Rubenstein, K.E. (1976) Fluorescent excitation transfer immunoassay. A general method for determination of antigens. J. Biol. Chem. 251, 4172-4178 16 Rodbard, D. (1978) Statistical estimation of the minimal detectable concentration (“Sensitivity”) for radioligand assays. Anal. B&hem. 90, I- 12 17 Combleet, P.J. and Goohman, N. (1979) Incorrect least-square regression coefficients in method analysis. Clin. Chem. 25, 432-438 18 Woodhead, J.S., Addison, G.M. and Hales, C.N. (1974)The immunoradiometric assay and related techniques. Br. Med. Bull. 30, 38-49 19 Nargessi, R.D., Landon, J., Pourfarzaneh, M. and Smith, D.S. (1978) Solid-phase fluoroimmunoassay of human albumin in biological fluids. Clin. Chim. Acta 89, 455-460 20 Barbour, H.M. (1976) Development of an enzyme immunoassay for human placental lactogen using labelled antibodies. J. Immunol. Methods 1I, 15-23