PRENATAL DIAGNOSIS, VOL. 1 1,427-435 (1 99 1 )


Department of Human Genetics, Klinikum der Chrbtian-Albrechts-Vniversitat Kiel, F.R.G.

SUMMARY Isoelectric focusing (IEF) of amniotic fluid a-fetoprotein (AFP) in thin-layer polyacrylamide gels containing 8 M urea followed by immunoblottingreveals at least nine bands, band I lying next to the cathode. Compared with 298 amniotic fluid samples from normal piregnancies, we found1 that the density of band V was increased in seven cases of fetal death. In 16 amniotic fluid samples from pregnancies with open neural tube defects (ONTD), band 'V disappeared or was markedly decreased. In seven cases with elevated AFP and positive atcetylcholinesterase: (AChE) due to contamination with fetal blood, no difference in pattern was observed compared with samples from normal pregnancies. It is suggested that IEF of AFP and subsequent immunoblotting are an apparently diagnostic test for ONTD and intrauterine fetal death (IUFD). KEY WORDS

Amniotic fluid a-Fetoprotein Isoelectric focusing Microheterogeneity Neural tube defect Intrauterine fetal death

INTRODUCTION a-Fe toprotein (AFP) measurement in amniotic fluid is a valuable and wellestablished tool in prenatal diagnosis, especially in testing for open neural tube defects (ONTD). Because of the possibility of false-positive and false-negative results, supplementary tests such as ultrasound examination and qualitative analysis of AChE are necessary to make a final diagnosis. Electrophoretic analysis of AFP has shown it to be heterogeneous, probably due to diifferences in its carbohydrate moiety (Smith and Kelleher, 1980). In recent studiies, AFP was separated by electrophoresis in gels containing concanavalin A (con A) into lectin-reactive and non-reactive fractions (Hindersson et al., 1979, Noergaard-Pedersen et al., 1980; Taketa et al., 1985; Toftager-Larsen and Noergaard-Pedersen, 1988). The percentage of con A non-reactive amniotic fluid AFP was found to be abnormally low in pregnancies with ONTD and other severe fetal abnormalities (Noergaard-Pedersen et al., 1980). Compared with quantitative AChE analysis, there was no significant difference in sensitivity for detecting fetal malformations, but more false-positive results were found for the con A analysis (Toftager-Larsen and Noergaard-Pedersen, 1988).Therefore separation of amniotic fluid AFP with higher resolution electrophoretic techniques was expected to improve the prenatal diagnosis of ONTD and other fetal abnormalities. In this paper we describe a new method for separating amniotic fluid AFP fraciions, using IEF in thin-layer polyacrylamide gels containing 8 M urea Addressee for correspondence: Kai Wiechen, Department of Human Genetics, Klinikum der Christian-Albrechts-UniversitltKiel, Schwanenweg24, D-2300 Kiel 1, F.R.G.

0197-385 1/9 1 /070427-09$05.00 0 1991 by John Wiley & Sons, Ltd.

Received 5 September 1990 Accepted23 February 1991



followed by sensitive enzyme-linked immunoblotting, and its application in prenatal diagnosis. MATERIAL AND METHODS Polyacrylamide gel

Acrylamide, N,K-methylenebisacrylamide(Bis) and Pharmalyte were obtained from Pharmacia-LKB. All other chemicalswere of analytical grade. Polyacrylamide IEF gels (T = 5.2 per cent, C =: 3-85per cent) consisted of 4.6ml stock solution (21.74g acrylamide and 0.87 g Bis/100 ml), 1-0ml Pharmalyte (pH 4 - 4 5 ) , and 9.61 g urea made up to 19.5ml with H,O. After degassing, polymerization was performed by adding 500 pl ammonium persulphate (15 mg/ml) and 15 p1 N,N,K,W-tetramethylethylenediamine(TEMED, Serva). IEF gels (230x 110 x 0.5 mm) were cast by the 'flap' technique (Radola, 1980) using parafilm spacers. Different kinds of polyester sheets (No. 146653,Desaga; Gel-Fix for Covers, Serva) were used to cover the lower and upper glass plates. Polymerization proceeded for 2 h. Because of possible carbamylation of amine and sulphydryl groups of proteins, the gel should be used the same day as it is prepared. Isoelectric focusing

The anolyte and catholyte were 0.04 M histidine, respectively. IEF was performed at 15°Cusing a 21 17 Multiphor I1 apparatus (Pharmacia-LKB). The interelectrode wick distance was 10cm. After prefocusing with 400 V for 20 min, 10 pl samples, diluted to 4 pg/ml AFP with H,O,were applied 2 cm from the anode, using an applicator strip (7 x 1 mm, Serva). Separation was performed at 500, 1000, and 1500V, 30 min each, and at 2000 and 2500 V, 20 min each. Power was limited to 30 W. Immunoblotting

After completion of IEF and removal of the polyester backing, focused proteins were transferred by electroblotting to nitrocellulose (NC)membranes (Sartorius, 0.45pm pore size), using a Trans-Blot Cell (Bio-Rad Laboratories). NC membranes were previously prewetted in transfer solution (0.7per cent acetic acid). Transfer was carried out at 70 V and 18°C for 2 h. After blotting, the membranes were blocked with 3 per cent bovine serum albumin (BSA, Biomol) dissolved in 20 m~ phosphate-buffered saline (PBS), pH 7.3,for at least 12 h at 8°C. The membranes were then exposed for 2 h to rabbit anti-human AFP IgG (Dako, A 008, lot 097 B) diluted 1: 500 in 100 ml of dilution buffer (PBS with 1 per cent BSA and 0.1 per cent Tween 20)followed by two 10 min washing steps with 0.1 per cent Tween 20 in PBS. The filter was then probed to swine anti-rabbit immunoglobulins conjugated with alkaline phosphatase (Dako, D 306)at a dilution of 1:2000.After incubation for 2 h, the membranes were washed twice in PBS with 0.1per cent Tween 20 (10 min), once with PBS (5 rnin), and once with 0.1 M diethanolamine buffer, pH 9.6 (5 min). Colour was finally developed by treating the membranes for 10 min with 4 mg of 5-bromo-4-chloro-3-indolylphosphate(BCIP, Sigma), 6 mg of Nitro Blue tetrazolium chloride (NBT, Serva), and 200 p1 of 1 M MgCl, in 60 ml of 0.1M



diethanolamine buffer, pH 9-6. Staining may be prolonged overniglht protected from light for very weak bands (Ramlau, 1987).The colour reaction was stopped by immersing the membranes in 5 per cent trichloracetic acid. The membranes were then extensively washed with distilled water, dried, and photographed. For treatment with monoclonal antibody, the blocked membrane was cut into 2 x 6 cm strips. The strips were incubated for 16 h in 20 pg of monoclonal anti-AFP (Dr Molter, AFD 046) in 2 ml of H,O and 8 ml of dilution buffer and then probed for 2 h to rabbit anti-mouse conjugated with peroxidase (Dako P 260) diluted 1:20010. Colour was developed with 3,3’-diaminobenzidine according to Ramlau ( I 987’). Densitometric evaluation

Stained membranes were treated with microscope immersion oil and scanned at 540 nm, using a Desaga CD60 densitometer. Samples

The samples were as follows: (1) 70 amniotic fluid samples from unaffected pregnancies with AFP levels less than 2.5 multiples of the normal median (MOM); (2) 230 amniotic fluid samples consecutively received in our laboratory with AFP levels less than 2.5 MOM; (3) 6 amniotic fluid samples from cases with intrauterine fetal death (IUFD) diagnosed after amniocentesis; and (4) 24 amniotic fluid samples with visible AChE bands. These included 8 bloodstained samples from pregnancies with normal outcome, which showed faint AChE bands in 7 cases and a dense AChE band in 1case; 16 samples associated with ONTD had dense AChE bands.

All amniotic fluid samples were stored at - 20°C until they were analysed. The AFP concentration was determined by rocket immunoelectrophoresis(NoergaardPedersen, 1972). AChE polyacrylamide gel electrophoresis was canied out as described by Haddow and Goldfine (1985). RESULTS The .AFP pattern in pregnancies with normal outcome obtained by polyacrylamide gel IEF in the presence of 8 M urea and subsequent immunoblotting is shown in Figures 1,2, and 5A. With this method, amniotic fluid AFP from nonmal pregnancies is represented by at least nine major bands which may be combined with some mort: or less pronounced minor bands. Since it was not possible to visualize the most anodal band in all cases, bands were numbered beginning with the most cathodal band. In order to establish this AFP IEF pattern, we analysed 70 samples from the 15th to the 22nd week of gestation from unaffected pregnancies with AFP levels below 2.5 MOM retrospectively. The monoclonal antibody detected tlhe same AFP pattern as the polyclonal antibody (not shown). Furthermore, we examined 230 amniotic fluid samples consecutively received in our laboratory from the 13th to the 32nd week of gestation with AFP concentrations



Figure 1. Amniotic fluid AFP pattern obtained by IEF in polyacrylamidegel (5.2 per cent T, 3.85 per Cent C )containing 8 M urea and 5 percent Pharmalyte,pH 4-6.5, followed by immunoblotting.The cathode is at the top. The separation distance was 10cm.Diluted samples were applied 2 cm from the anode. The 0.5 mm thick gel was run for 3000 V h. All samples from normal pregnancies show a uniform pattern

Figure 2. Detail of Figure 1. IEF immunoblottingpattern of amniotic fluid AFP on polyacrylamidegels containing 8 M urea, pH 4-6.5, using diluted samplesfrom normal pregnancies. Arrows indicate bands I and V

below 2.5MoM. All of these samples showed the same uniform pattern just described, except for two samples, in which we observed a quite different pattern (Figures 3 and 5C) with the density of band V increased and some additional bands


43 1

Figure 3. IEF immunoblottingpattern of amnioticfluid AFP on polyacrylamide gels cont aining 8 M urea, pH 4 - 6 5 , using diluted samples. Arrows indicate band V. Lanes 1 and 2 are samples from normal pregnancies. An amniotic fluid sample from a pregnancy with IUFD (3) shows that the density of band V is increased and there are some additional bands especiallyin the acidic region of the gel

especially in the acidic region of the gel. IUFD was diagnosed in both cases. The same AFP pattern with V as the major component was found in six additional cases of IUFD. Dense AChE bands were found in all eight cases of IUFD. Twenty-four amniotic fluid samples with dense positive or faintly positive AChE bands were then analysed. These included eight bloodstained samples from unaffected pregnancies, five of which had normal and three slighitly elevated amni,oticfluid AFP. Sixteen samples associated with ONTD (Table 1) had significantly elevated AFP concentrations. In all samples with a visible AChE,band due to fetal iblood contamination, we found AFP patterns indistinguishable from those of normal samples (Figure 5B). Different but constant patterns, however, were found in all 16 samples of pregnancies associated with ONTD (Figures 4 and 5D). In all samples from cases with ONTD, bands I and V were drastically decreased or the AFP concentrations of these fractions were below the detection limit of'this enzyme immunoassay.The acidic bands were more intense than those in the normal pattern. DISCUSSI 0N AFP is an a-1 globulin synthesized mainly by the fetal yolk sac and liver (Gitlin and Boesman, 1967). Increased AFP levels in amniotic fluid are found in ONTD and in a number of other fetal malformations, including ventral wall defects and congenital nephirosis (Brock, 1976). However, amniotic fluid AFP is also elevateld in samples with '[UFD, multiple pregnancies, and most frequently in samples with fetal blood



Table 1. Amniotic fluid AFP levels in pregnancies associated with open neural tube defects Case No. 11 2 3 4 5 6 7

8t 9t 10 11 12 13 14 15 16

Diagnosis Anencepha1y Anencephaly Anencephaly Anencephaly Anencephaly Anencephaly Spina bifida Spina bifida Spina bifida Anencephaly Anencephaly Anencepha1y Anencephaly Anencephaly Anencephaly Anencephaly

Amniotic fluid AFP (MOM)


4.8 5.0 3.8 17.0 4.5 12.7 4.3 3.3 4.2 6-0 5.1 5.3

19 18 16 21 17 17 16 17

15.0 25.5 21.0 4.0


17 13 19 17 24 29 16 18

*Anencephalycombined with encephalocele. tspina bifida combind with trisomy 18.

Figure 4. IEF immunoblottingpattern of amnioticfluid AFPmpolyacrylamide gels containing8 M urea, pH4-6-5, using diluted samples. Arrows indicate bands I and V. Lanes 1 and 2 are samples from pregnancies with ONTD. Bands I and V are missing; the acidic bands are more pronounced than in the pattern from a normal pregnancy (3)





Figure 5. Comparison of densitometric profiles of AFP IEF patterns. Nitrocellulose membranes were treated with microscope immersion oil and scanned at 540 nm. The shaded areas indicate bands I and V. Peak in A are numbered from the cathode. Amniotic fluid samplesare from (A) a normal prenancy,(B) a normal pregnancy with a visible AChE band due to contaminationwith blood, (C) acasa with IUFD, and (D) a case with ONTD

contamination (Seppala and Rouslahti, 1972; Ward and Stewart, 1974).Therefore supplementary tests are necessary to confirm the diagnosis. Presently, qualitative analysis of AChE is widely used as a secondary test for ONTD, but sometimes falsepositive results occur in other conditions, such as IUFD and contamination of amniotic fluid by blood (Seller ef al., 1980). In this paper we have described a new method for separating amniotic fluid AFP by IEF in the presence of 8 M urea and subsequent immunoblotting. This system comlbines the resolution of IEF with the sensitivity and specificity of an enzymelinked immunoassay. The use of untreated polyester sheets as gel backing allows simple removal for blotting purposes. The electroblotting technique is more efficient and less time-consuming than capillary blotting. Because conventional polyacrylamide gels (T = 5 per cent, C = 3 per cent) tend to be sticky and adhere to nitrocellulose membranes, we used more robust gels (T = 5.2 per cent, C = 3.85 per cent) for electroblotting, which can easily be removed from nitrocellulose prior to the immunoassay. Isoelectric focusing in narrow pH gradients is a technique with a very high resolution and protein microheterogeneity is frequently observed. The origin of the AFlP microheterogeneity seems to lie in differencesin the carbohydrate moiety of the molecule (Smith and Kelleher, 1980). We were able to separate nine major AFP bands and a variable number of minor bands or doubled bands in (amnioticfluid samples from normal pregnancies. This banding pattern remained constant for the analysed interval from the 13th to the 32nd week of gestation. In cases associated with ONTD and intrauterine fetal death, we observed two altered patterns that were specific and constant for the particular complication of pregnancy. However, no variation in AFP pattern, compared with other normal pregnancies, was observed when bloodstained amniotic fluid samples were analysed.



Unfortunately, our amniotic fluid samples did not include cases with other fetal malformations, such as ventral wall defects or congenital nephrosis. Thus, we do not know yet whether AFP IEF patterns are altered in these conditions. Although the AFP pattern detected by the monoclonal antibody was in agreement with the bands recognized by the monospecificpolyclonal antibody, it may be possible to identify single AFP fractions with different monoclonal antibodies. In this respect, immunodetection of AFP fractions I and V with monoclonal antibodies might improve diagnostic tests for ONTD and IUFD. Our results suggest that IEF of amniotic fluid AFP in the presence of 8 M urea followed by immunoblotting may be used as a diagnostic test for ONTD and fetal death. This qualitative analysis of AFP is especially capable of distinguishing unaffected pregnancies from those associated with ONTD in cases where AChE gel electrophoresis is not decisive, for example in amniotic fluid samples contaminated with blood.


We wish to thank M. Krussek an'd C. P. Blohm for photographs. REFERENCES Brock, D.J. (1976). Mechanisms by which amniotic fluid alpha-fetoprotein may be increased in fetal abnormalities, Lancet, ii,345-346. Gitlin, D., Boesman, M. (1967). Sites of serum a-fetoprotein synthesisin the human and in the rat, J. Clin. Invest., 46, 1010-1016. Haddow, J.E., Goldfine, C. (1985). The evolving role of amniotic fluid acetylcholinesterase analysis for identifying open neural tube defects during the second trimester. In: Mizejewski, G.J., Porter, I.H. (Eds). Alpha-fetoprotein and Congenital Disorders, San Diego, CA: Academic Press, 215-237. Hindersson, P., Toftager-Larsen, K., Noergaard-Pedersen, B. (1979). Concanavalin A reactivity of amniotic fluid alphafetoproteinin the diagnosis of neural tube defects, Lancet, ii, 906. Noergaard-Pedersen, B. (1 972). Punfication and sensitive immunoelectrophoreticaldetection and quantification of human alpha-1-fetoprotein,Cfin.Chim. Acta, 38,163-170. Noergaard-Pedersen, B., Toftager-Larsen, K., Philip, J., Hindersson, P. (1980). Concanavalin A reactivity pattern of human amniotic fluid AFP examined by crossed affino-immunoelectrophoresis. A definite test for neural tube defect?, Clin. Genet., 17, 35$-362. Radola, B.J. (1980). Ultrathin-layer iiioelectric focusingin 50-100 pm polyacrylamide gels on silanized glass plates or polyester films, Electrophoresis, 1,43-56. Ramlau, J. (1987). Use of secondary antibodiesfor visualisation of bound primary reagents in blotting procedures, Electrophoresis, 8,398402. Seller, M.J., Cole, K.J., Fensom, A.H., Polani, P.E. (1980). Amniotic fluid acetylcholinesterase and prenatal diagnosis, Br. J. Obstet. Gynaecol., 87,501-505. Seppzil5, M., Rouslahti, E. (1972). Alpha-fetoprotein in abortion, Br. Med. J., 4,501-505. Smith, C.J.P., Kelleher, P.C. (1980). Alpha-fetoprotein molecular heterogeneity, Biochim. Biophys. Acta, 605,1. Taketa, K., Ichikawa, E., Taga, H., I-Iirai, H. (1985). Antibody-affinity blotting, a sensitive technique for the detection of a-fetoprotein separated by lectin affinity electrophoresis in agarose gels, Electrophoresis, 6,492497.



Toftager-Larsen, K., Noergaard-Pedersen,B. (1988).Con A non-reactivefractionsof human amniotic fluid alpha-fetoproteinin prenatal diagnosis of fetal neural tube defects and fetal abtdominal wall defects, Clin. Genet., 33,220-227. Ward, A.M., Stewart, C.R. (1974). False-positive results in antenatal diagnosis of neuraltube disorders, Lancet, ii, 345-346.

Isoelectric focusing pattern of human amniotic fluid alpha-fetoprotein.

Isoelectric focusing (IEF) of amniotic fluid alpha-fetoprotein (AFP) in thin-layer polyacrylamide gels containing 8 M urea followed by immunoblotting ...
606KB Sizes 0 Downloads 0 Views