Carbohydrate heterogeneity of human alpha-fetoprotein in pregnancy. Methodological and clinical studies

The current status and development of biochemical analyses on amniotic fluid in prenatal diagnosis of congenital malformations is shortly summarized. Based on reports on heterogeneity in the carbohydrate content of alphafetoprotein (AFP) a method for electrophoretic separation of heterogeneous fractions of AFP was developed. In principle the method was first described to predict the separation of glycoproteins on affinity chromatography columns. In the analysis AFP is first electrophoresed in an agarose gel containing free concanavalin A (con A). Thereby, AFP is separated into two fractions (with and without affinity to con A). By further electrophoresis--perpendicular to the first--into a gel containing monospecific antibody to AFP two precipitates are formed, where the relative size of the fraction without affinity to con A is determined (con A non-reactive AFP). To obtain the smallest possible analytical variation the method was modified by introduction of a line gel, which not only increased the precision, but also made the reading of the gels considerably easier. Also, casting of the antibody containing gel between two glass plates to obtain an equal thickness further reduced the variation. On each second dimension plate three patient samples and a control sample with a non-reactive fraction of AFP corresponding to the limit between samples from normal pregnancies and from pregnancies with fetal malformations were run. Using these modifications a constant coefficient of variation of 5 to 6% over a six-year period was obtained. The investigations showed that at least 50% of AFP from the fetal yolk sac (where the primary production takes place) consists of the non-reactive fraction, whereas only a few per cent of AFP from the fetal liver (where the secondary and major production takes place from 9-10 weeks) is of the non-reactive type. Correlated to the development of the fetal anatomy and physiology with advancing gestation it was expected that AFP in very early amniotic fluid mainly would be of the yolk sac type, but that a decrease in the non-reactive fraction could be expected by means of an elimination of the yolk sac AFP through the fetal gut and a partial replacement by fetal liver AFP (from the fetal liver) through fetal urine. In agreement with this theory a significant decrease in the fraction of AFP non-reactive with con A from high values (30-40%) in amniotic fluid samples from early pregnancy (8-12 weeks) to low values (0-3%) in amniotic fluids from late second trimester was found.(ABSTRACT TRUNCATED AT 400 WORDS)