CLIN. CHEM. 29/1, 115-119 (1983)

Use of Monoclonal Antibodies to Detect Human Placental Alkaline Phosphatase Guy De Groote,1 Peter De Waele,1 Andre Van de Voorde,1 Mark De Broe,2 and Walter Fiers’ Convenient, sensitive, and specific solid-phase immunoassays involving monoclonal antibody are described for the determination of human placental alkaline phosphatase

body. This has allowed us to establish a monospecific immunoassay to detect hPLAP from various sources. The method could be particularly useful in the evaluation of

(hPLAP). An endogenous enzyme immunoassay combined

alkaline phosphatase in human sera during pregnancy in monitoring of seropositive cancer patients.

the specificity of the immunological and the enzymatic reactions. Alternatively, a solid-phase “sandwich” radioimmunoassay involving immobilized polyclonal rabbit antihPLAP in combination with iodinated monoclonal antibody provided some additional advantages. Both tests can be

used to detect hPLAP from various sources, e.g., in human sera during pregnancy

or as a tumor marker.

The radio-

immunoassay detected an increase in hPLAP at nine weeks of gestation. We discuss the use of monoclonal antibodies for

the differentiation

of different alkaline phosphatase

isoen-

by electrophoresis on starch gel.

zyme types

AddItionalKeyphrases:enzyme immunoassay immunoassay isoenzymes pregnancy phoresis, starch gel tumor-related antigens ‘

.

radioelectrocancer

alkaline phosphatase, hPLAP (EC produced by the microvilli of the syncytiotrophoblast (1), but is also expressed as an oncodevelopmental protein in various cancer tissues and can be found in the sera of some cancer patients (2). As such, ithas been proposed as a potential marker suitable for monitoring Human

placental

3.1.3.1), is normally

cancer

patients.

A major difficulty in the detection of hPLAP

stems from the fact that other groups of isoenzymes (bone, liver, and intestinal types) may also be present in the sera. These isoenzymes are thought to be coded by at least three different genes (3). Moreover, the intestinal and the placental enzymes share some antigenic determinants and have

several tryptic peptides in common. Consequently, a variety heat inactivation (4), sensitivity to inhibitors such as phenylalanine (5) or bromotetramisole (6), differential migration in several electrophoretic systems (7), and several immunological assays (8, 9)-have been proposed to detect hPLAP and to differentiate between hPLAP and other alkaline phosphatases present in human serum. Immunological techniques are severely hampered by the cross reactivity of most polyclonal antibodies to hPLAP with human intestinal alkaline phosphatase, and so far only intensive adsorption of the immune sera has successfully eliminated this cross reactivity (10). We have overcome this problem by isolating a hybridoma cell line that produces an hPLAP-specific monoclonal anti-

of techniques-including

‘Department L.edeganckstraat

of Molecuiar Biology, State University of Ghent, 35, B-9000 Ghent, Belgium. of Nephrology-Hypertension, University of Ant-

2Department werp, Belgium. Received July 26, 1982; accepted Sept. 22, 1982.

and

Materials and Methods hPLAP was prepared and purified as described by Lehmann (11). Human intestinal and liver alkaline phosphatases present in butanol extracts were dialyzed according to Sussman et al. (12). The isolation of the hybridoma cell line that secretes monoclomd antibody against hPLAP (clone E6) and the production of polyclonal rabbit anti-hPLAP antiserum will be described elsewhere (De Waele et al., in preparation). Sera from women at various stages of pregnancy were obtained from clinical laboratories and stored at

-20 #{176}C until use. Preparation

of solid-phase

antibody.

Polystyrene

beads

(5

mm in diameter; Precision Plastic Ball Co., Chicago, IL) were coated with antibodies by a modification of the te.hmque described by Hendry and Herreman (13). The beads were treated overnight at room temperature with 3.5 mol/L HC1. After being washed twice with doubly-distilled water they were left for 1 mm in bicarbonate buffer (0.3 mol/L, pH 9.5), washed again with phosphate-buffered saline (PBS: 10 g/L NaCl, 0.25 g/L KC1, 1.43 g/L Na2HPO4, and 0.25 g/L KH2PO4), and incubated for 2 h at room temperature with glutaraldehyde in a final concentration of 20 g/L. After two

washes with doubly-distilled water, the beads were incubated for 18 h at room temperature in PBS containing an optimal dilution of antibody (generally 1000-fold). As antibody sources, whole rabbit serum, purified rabbit IgG, further

hybridoma cell culture supernate, ascites fluid from hybridoma cells grown in Balb/c mice (as in the experiments shown later in Figures 1, 2, and 4) or in nude mice, and purified monoclonal antibody can be used. The beads were then left for 1 h at room temperature in a blocking solution containing, per liter, 2.5 g of bovine serum albumin, 10 mL of fetal calf serum, and 2.5 mL of control rabbit serum. After repeated washings with PBS containing 0.5 mL of Tween 80 per liter, they were stored dry at 4 #{176}C until use. /ZPLA.P endogenous enzyme immunoassay. Our undiluted polyclonal rabbit antiserum neutralized only 10% of the

activity and our monoclonal antiserum activity at all; therefore, we were able to use these properties to set up a solid-phase enzyme immunoassay based on the endogenous enzymatic activity of the hPLAP-enzymatic

had no neutralizing antigen.

In this assay, beads coated with polyclonal or monoclonal antibody were incubated overnight at room temperature with 200 L of the samples to be tested for hPLAP activity. After repeated washing, the beads were transferred to fresh CLINICALCHEMISTRY,Vol. 29, No. 1, 1983

115

tubes and the remaining

endogenous alkaline phosphatase was determined by incubation with 500 L of p-nitrophenyl phosphate, 4.5 mmolJL of diethanolamine buffer (14) for 1 h at 37 #{176}C. The enzymatic reaction was stopped by the addition of 500 L of 1 mol/L NaOH and the absorbance was measured at 410 nm. One unit (U) of enzyme activity is defined as the liberation of 1 prnol of phosphate from pnitrophenyl phosphate (e 14 600) per minute at 37 #{176}C when incubated in a solution containing 4.5 mmol of pnitrophenyl phosphate per liter of diethanolamine buffer (110 g/L).

a 0.6

=

1ZPLAP

solid-phase

“sandwich”

RIA.

Beads

coated

with

polyclonal rabbit anti-hPLAP were incubated with test samples containing hPLAP as described, washed, and then incubated again for 1 h at room temperature with 200 tL of ‘251-labeledmonoclonal antibody that had been iodinated by the Chloramine T method (15) and which was used in a concentration of 0.75 mCiIL. After repeated washings the remaining radioactivity was counted with a gamma counter (LKB Productor, Bromma, Sweden). In the same assay we were also able to measure the residual enzymatic activity as described in the previous section.

0.4

E

C 0

,(

0.2

Starch gel electrophoresis. Alkaline phosphatase isoenzymes in human sera were differentiated by electrophoresis

in starch gels according to Poulik (16). The procedure involves a flat-bed 120 g/L starch gel (13 x 15 cm), cast in a pH 8.65 buffer consisting of 76 mmol of Tris and 5 mmol of citric acid per liter. Samples are adsorbed on 9 x 4 mm Whatman 3 MM papers that are inserted into precut slots at the start. The running buffer (pH 8.86) consists of 0.3 mol of boric acid and 0.05 mol of NaOH per liter. Gels are run at 160 V (30 mA) for 2 to 3 h. Alkaline phosphatase isoenzymes were made visible by using f3-naphthyl phosphate (Aldrich Chemical, Beerse, Belgium) as a substrate and Fast Blue HR salt (Difco Laboratories, Detroit, MI 48232) for coupling. i.-p-Bromotetramizole (Aldrich) at concentrations of 5 and 0.01 mmolJL was used to differentiate between placental and intestinal isoenzyme activity (6).

-

b 0.6

-

0.4

E C -.:

0.2

Results Specificity of the Monoclonal Antiserum, as Determined by the Endogenous Enzyme Immunoassay The glutaraldehyde

technique

yielded

a reproducible,

stable coating on the polystyrene beads-better results than with a coating procedure involving passive absorption in a low-concentration phosphate buffer (10 mmol/L) or a barbital/citric acid/boric acidimonopotassium phosphate buffer (3 mmol/L each) at pH 7.0 (data not shown). The fact that we were still able to detect enzymatic

activity in our solid-phase immunoassay (see Materials and Methods) permitted us to check the cross reactivity of polyclonal or monoclonal antisera for human liver AP and human intestinal AP. Cross reactivity was observed only with polyclonal rabbit antiserum, none with the monoclonal antibody (Figure 1). In no instance could we demonstrate immunobinding of human-liver alkaline phosphatase, either with the polyclonal or with the monoclonal antibody. At an equal input of enzymatic activity (expressed in U/L) the amount of activity retained on the beads coated with polyclonal antiserum was at least 16-fold greater for hPLAP than for intestinal alkaline phosphatase. In the solid-phase assay with monoclonal antibody bound to the beads, addition of a constant amount of the monoclonal antibody (in excess, as compared with the amount bound on the beads) to the hPLAP-containing sample during incubation neutralized the antigen binding by more than 95%. Adding mono116 CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983

-

-I

1

3

6

10

dilution

factor

Fig. 1. Endogenous enzyme immunoassay studied by coating beads

with either monoclonal (a) or polyclonal (b) anti-hPLAP sera Lines ShOW the enzyme activityretained by various dilutions of hPLAP, 38 U/L (C #{149}), human liver alkaline phosphatase,33 U/L (A-A), and human intestinal alkaline phosphatase,123 U/L (#{149} U). Open symbols represent results of a similar assay in which soluble monoclonal anti-hPLAP was used

nal antibody and an hPLAP preparation to solid-phase bound polyclonal antibody neutralized only 50 to 60% of the binding. This indicated that the epitope recognized by the monoclonal serum only partly overlaps with one of those represented in the polyclonal serum. Solid-Phase

Sandwich

Radioimmunoassay

To select an assay with maximal specificity and optimal sensitivity, we tested several combinations of monoclonal and polyclonal antibodies in a solid-phase sandwich-type radioimmunoassay (see Materials and Methods). The results

a +

1

2

3

4

5

:

6

11

12

b +

C

4

0

0.0!.

017

0.36

32

1.1

96

709

866

hPLAP. U/L

Fig. 2. Comparison ofsolid-phase

AlA and EIA The solid-phase AlA involved different combinations with polyclonal (Rb)or monoclonal (M) antibodyas solid phase and raoloiodinatedantibodies.Results expressed as cpm are shown on the left ordinate:#{149}, Rb/M (polyclonalsolid phase/monoclonal label); 0, Rb/Rb; A, M/M; & MAb. The right ordinate shows the enzymaticactivityofthesolid-phaseendogenousenzymeImmunoassaywith monoctonalantibody(#{149}#{149}) overthe same rangeof hPLAP concentrations

(Figure 2) indicate that a combination of a solid-phasebound polyclonal antibody with an iodinated monoclonal

antibody provided both the advantage of the specificity of the monoclonal antibody and the optimal radioactive signal over a wide range of dilutions of hPLAP. When monoclonal antibody was used as both first and second antibody, the sensitivity was considerably less than in the assay system just described, which also gave a better signal-to-noise ratio over a wider range of enzyme dilutions. The sensitivity of the selected RIA combination (polyclonal solid-phase, labeled monoclonal) was comparable with, or slightly better

than, that of the endogenous enzyme immu-

noassay method. Both methods enabled the detection of hPLAP in serum at 0.12 U/L. The between-assay variation (CV) was 11.4% and 5.1% for sera containing hPLAP of 5 and 30 UIL, respectively. The within-assay CVs were 4.1, 6.0, and 7.9% for respective hPLAP activities of 25.5, 12.8, and 6.4 U/L.

Reactivity of the Monoclonal Antibody with Different Alkaline Phosphatase Phenotypes As can be seen from Figure

3a, the monoclonal

antibody

does not interact with serum alkaline phosphatases of nonplacental origin. When polyclonal rabbit antiserum against hPLAP is incubated with samples containing both liver and intestinal AP, only the latter component disappears from the mixture (Figure 3a, lane 12), while incubation with monoclonal antibody (Figure 3a, lane 11) does not alter the mobility patterns as compared with the control. The cross reactivity of the polyclonal antiserum is to be expected because both isoenzymes have several tryptic peptides in common (3) and hence they share some antigenic determinants. Incubation of monoclonal antiserum with each of the three most abundant allelic types of placental alkaline phosphatase resulted in precipitation of the major band at

the site of application,

indicating

that

1ST

14

the monoclonal

Fig. 3. A: Alkaline phosphatase (AP) isoenzyme pattern of human serum in starch gel electrophoresis incubated with monoclonal or polyclonal anti-hPLAP serum slot 1 serum of patient with hepatitis alkaline phosphatase(3of the livertype(total AP activity,380 U/L). 2 same as slot 1.but incubated with monoclonalmouseanti-hPLAP. 3 same as slot 1, but incubated with polyclonal rabbitanti-hPLAP. 4 serum of patientwith Paget’sdisease (totalAP activity350 UIL);almost all AP is of bone type. 6

same as slot same as slot

7

immunoglobulinG-associatedAP in the serum of a patientwith ulcerative

5

4, 4,

but incubated with monoctonal mouse anti-hPLAP. but incubated with polyclonal rabbit anti-hPLAP.

colitis (total AP activity 405 U/L). 8 same as slot 7, but incubated with monoclonal mouse anti-tiPLAP. 9 same as slot 7, but incubatedwith polyclonal rabbit anti-hPLAP. 10 serum of patient with terminal renal failure treated with intermittenthemodialysis;the alkaline phosphatase is of the liver and intestinal type (total AP activity 285 U/L). 11 same as slot 10,but incubatedwith monoclonal mouse anti-hPLAP. 12 same as slot 10, butincubatedwith polyclonal rabbit anti-hPLAP. -. denotes application line; +, anode. AP-activity was made oisibleby using 13naphthylphosphate as substrate and Fast Blue AR salt as coupling agent.

B: Starch gel electrophoresis of three different types (F, FS or S) of hPLAP (18 mU/slot) incubated overnight at 37#{176}C in the absence or presenceof 20 L ofundiluted monoclonalantiserum (titer 10) before application Slots 1. 4, and 7: Control human liver alk. phosphatase 2: hPLAP, typeF 3: same as slot 2,but incubatedwith antiserum 5: hPLAP, type FS 6: same as slot 5, but incubated with antiserum 8: hPLAP, type S 9: same as slot 8,butincubated withantiserum

antiserum

recognized

typic variants, Detection

an epitope common to all three pheno-

albeit with a slightly

of hPLAP

in Serum

different affinity.

during

Pregnancy

Individual serum samples from 191 women at various stages of pregnancy were tested with the solid phase sandwich RIA. Figure 4 shows the amount of hPLAP (expressed as cpm) in relation to the duration of pregnancy. hPLAP markedly increases from about 11 weeks after the last menstruation (i.e., nine weeks of gestation). During the second trimester, all hPLAP values are usually above normal (i.e., nonpregnant) values, reaching about 25 U/L at the end of the second trimester and more than 100 U/L at full-term. Some samples show aberrant values that are either too high or too low for their gestational period. In samples with a high-for-date hPLAP value, unconjugated estriol and

CLINICALCHEMISTRY, Vol. 29, No. 1, 1983

117

weeksof pregnancy

Fig. 4. Application of the solid-phase AlA to the analysis of clinical samples: hPLAP concentrationas a function of duration of pregnancy in 191 serum samples from pregnant women The horizontal dotted line indicates the mean + 2 SD as calculated from measurements on a control group of 59 healthy, nonpregnant adults. Vertical bars indicatemean and range. human placental lactogen values were also increased in those individuals, so that probably the duration of those pregnancies were underestimated. In contrast, samples with

hPLAP activities that were low-for-date (as observed in some third-trimester samples) had normal hormonal values.

Discussion The endogenous ELA proved to be a sensitive and simple technique for the specific determination of hPLAP. By combining the specificity of both the monoclonal antibody and the endogenous enzyme reaction, the endogenous ETA can be a valuable addition to the spectrum of available solidphase immunoassays, especially when monospecific, nonneutralizing antibodies are used. This assay eliminates the need to use radioactivity, but has the disadvantage that longer (1 h at 37 #{176}C) incubation times are required to attain a sufficient enzymatic reaction at low concentrations of hPLAP. If radioactivity can be used, the sandwich solidphase RIA is recommended:

the linear part of the standard

curve for this method covers a substantially wider range of hPLAP dilutions, and it may also be possible to detect those hPLAP molecules that have lost their enzymatic activity. We tested several possible combinations of antisera in our solid-phase RIA. For every combination the radioactive signal produced by the iodinated monoclonal antibody was far superior to that obtained with iodinated polyclonal (not immuno-adsorbed) antibody. The sensitivity of the assay decreased considerably when monoclonal or polyclonal antibody was used as both solid phase and labeled antibody because then the same epitope or epitopes were used. The combination of monoclonal solid phase/labeled poly118 CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983

clonal antibody resulted in a disappointingly low level of radioactivity and sensitivity although one might have expected that lower background values would allow a better detection limit. Despite this observed disadvantage, the signal-to-noise ratio at 29 U of hPLAP per liter is 3.1, as compared with 8.3 for the optimal combination. Our monoclonal antibody E6 apparently recognizes all the major hPLAP variants tested so far, although differences in specific affinities to different phenotypes seem to occur. A similar variation in specific affinity for phenotypic variants of monoclonal antibodies has been described (17), and monoclonal antibodies recognizing only one phenotype have been reported (T. Stigbrand and D. L. Millan, personal communication). The use of the solid-phase RIA or ETA for detecting hPLAP in the sera of pregnant women may have clinical applications, e.g., during pre-eclampsia. A comparison of the different methods described so far for detection of hPLAP in sera is hampered by the wide range of methods used (different substrates or incubation conditions, different buffer systems, and often different ways of expressing resultse.g., as a percentage of total AP activity, as U/L, in KingArmstrong units, placental isoenzyme units, cpm, absorbance, and so on). By our solid-phase RIA method, the increase of hPLAP above normal values is detectable at about 11 weeks after the last menstruation. From 17 weeks on, all values are at or above the normal limit (i.e., the mean value observed in normal nonpregnant adults plus two standard deviations). By the criterion of heat resistance (on exposure to 56 or 65 #{176}C for 30 mm) of the isoenzyme, hPLAP reportedly is not detectable before 20 weeks (6) or 30 weeks of gestation (2, 18, 19). Romslo et al.(5) and Fishman et al. (20), however, using a method of selective heat inactivation and incubation conditions followed by mathematical treatment of the data, were able to limit the lag phase to between five and six weeks. The use of phenylalanine inhibition has proven even less satisfactory than heat inactivation as an identification test for hPLAP (6). Inactivation of hPLAP by bromotetramisole is relatively more specific and reveals an increased activity after 17 to 20 weeks of pregnancy (7). After comparing heat inactivation, bromotetramisole inhibition, and an immunological precipitation assay, Bayer et al.(21) suggested that this technique is suitable for routine determination of hPLAP in serum samples. Immunological techniques that have been used to quantifr hPLAP include neutralization, precipitation, RIA, and enzyme immunoassay with polyclonal rabbit anti-alkaline phosphatase. These sera, unless extensively absorbed, cross react heavily with the intestinal enzyme. Using a precipitation method, Sussman et al. (8) demonstrated the presence of h.PLAP from the 12th week after the last menstruation. Lehmann (22), however, using a similar assay combined with heat inactivation, was not able to demonstrate a deviation from normal before the 20th week of gestation. Other sensitive enzyme immunoassay and RIA techniques have been described but the authors provide insufficient data on serum values during pregnancy (9), although the sensitivity reported seems comparable with that described here. In conclusion, the monoclonal anti-hPLAP antibody provides a stable, highly specific, constant reagent that does not interfere with the enzyme activity. It can be used in a solidphase RLA in combination with a polyclonal antibody or perhaps also with another monoclonal antibody in the first phase. Finally, the use of monoclonal antibodies as a reagent for isoenzyme immunoelectrophoresis allows intestinal and placental isoen.zymes to be differentiated.

We thank ProfessorJ. Urbain, J.-D. Franssen, and P. Henrion for their help with the production of hybridomas and Dr. M. Dhont for collection of the serum samples. This research was supported by a grant from the Kankerfonds Of the Algemene Spaar- en Lijfrente-

kas (ASLK) of Belgium. PDW holds a fellowship of the Instituut bar bevordering van het Wetenschappelijk Onderzoek in Nijverheid en Landbouw,

Belgium.

References 1. Messer RH. Heat-stable alkaline phosphatase as an index of placental function. Am J Obstet Gynecol 98, 459-465 (1967). 2. Nathanson L, Fishman WH. New observations on the Regan isoenzyme of alkaline phosphatase in cancer patients. Cancer 27, 1388-1397 (1971). 3. McKenna M, Hamilton TA, Sussman RH. Comparison of human alkaline phosphatase isoenzymes. Structural evidence for three protein classes. Bwchem J 187, 67-73 (1979). 4. Fishman WH, Bardawil WA, Habib HG, et al. The placental isoenzyme of alkaline phosphatase in sera of normal frequency. Am J Clin Pat/wi 57, 65-74 (1972). 5. Romslo I, Sagen M, Haram K. Serum alkaline phosphatase pregnancy: A comparative study of total L-phenylalanine-sensitive

in

and heat-stable alkaline phosphatase at 56#{176}C and 65 #{176}C in normal pregnancy. Acta Obstet Gynecol Scand 54, 437-442 (1979). 6. Van Belle H, De Broe ME, Wieme RJ. i-p-Bmmotetramisole, a new reagent for use in measuring placental or intestinal isoenzymes of alkaline phosphatase in human serum. Clin Chem 23, 454-459 (1977). 7. Fishman WH. Perspectives on alkaline phosphatase

isoenzymes.

Eur J Med 56, 617-650 (1974). 8. Sussman RH, Bowman M, Lewis JL. Placental alkaline phosphatase in maternal cy. Nature (London)

9.

Millan placental

serum 218,

during

359-360

normal

and abnormal

pregnan-

(1968).

JL, Stigbrand T. Sandwich enzyme immunoassay for alkaline phosphatase. Clin Chem 27, 2014-2018 (1981).

10. Lehmann F-G. Preparation

of monospecific

antisera

for immu-

noassay of human placental phatase. J Immunol Methods

(Began) and intestinal 36, 137-148 (1980).

alkaline

phos-

II. Lehmann F-G. Differentiation of human alkaline phosphatases by lectin binding affinity. Kim Wochenschr 58, 947 (1981). 12. Sussman RH, Small PA, Cotlove E. Human alkaline phosphatase. Immunological identification of organ-specificisoenzymes. J Biol Chem 243, 160-166(1968). 13. Hendry RM, Herreman JE. Immobilization of antibodies on nylon for use in enzyme-linked imniunoassay. J Immunoi Methods 35, 285-296 (1980). 14. Voller A, Bidwell JE, Berlett A. Enzyme immunoassays in diagnostic medicine. Theory and practice. Bull WHO 53, 35-65 (1976). 15. Cameron CH, Combridge

BS, Howell

DR, Barbara JAJ. A

sensitive immunoradiometric assay for the detection of hepatitis B surface antigen. J Virol Methods 1, 311-323 (1980). 16. Poulik MD. Starch gel electrophoresis in a discontinuous system of buffers. Nature (London) 180, 1477-1479, (1957). 17. Slaughter CA, CoseoMC, Camcro MR, Harris H. Detection of enzyme polymorphism by using monoclonal antibodies. Proc Nati Acad Sci USA 78, 1125-1128 (1981). 18. Curzen P, Morris I. Heat-stable alkaline phosphatase in maternal serum. J Obstet Gynaecol Br Comm.onw 75, 15 1-157 (1968). 19. Hunter JR, Crog MR, Pinkerton JHM, Johnston H. Serum placental alkaline phosphatase in normal pregnancy and preeclampsia. Obstet Gynecol 36, 536-546 (1970). 20. Fishman WH, Anstiss CL, Pirnik MP, Driscoll SG. The exponential growth curve for the placental isoenzyme of alkaline phosphatase in sera of normal and diabetic pregnancies. Am J Clin Pat/wI 60, 353-358 (1973). 21. Bayer PM, Gabl F, Knoth E. Aktivitatsbestimmung desplacentalen Isoenzyms der alkalischen Phosphatase im serum: Em Methoden Vergleich. J Clin Chem Clin Biochem 17, 605-607 (1979). 22. Lehmann F-G. Immunological methods for human placental phosphatase (Began isoenzyme). Clin Chim Acts 65, 271-282 (1975).

CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983

119