Production and Evaluation of a Two-Site Immunoradiometric Assay for Alpha-1-Foetoprotein Using Monoclonal Antibodies

Alpha-l-foetoprotein (AFP) is a glycoprotein present at high concentration in foetal serum until birth, with residual low levels present throughout adult life. The most frequent cause of elevated AFP concentration in adults is pregnancy, during which foetal AFP passes across the placenta and the amnion into the maternal circulation. The concentration of maternal serum AFP (MS-AFP) is related to the gestational age, and to the clinical status of the foetus and placenta. Elevated MS-AFP is a marker of several clinical entities, of which the most important and frequent are neural tube defects. In view of this, since 1975, the entire mid-trimester pregnant population of the West of Scotland has been offered MS-AFP screening and at the start of this project the West of Scotland maternal-serum screening programme used a RadioImmunoAssay (RIA) technique with Iodine labelled AFP and sheep anti-human AFP antiserum. Separation was effected by 16-hour incubation in polyethylene glycol 6000. 30,000 samples per year were being assayed and the workload was increasing. Disadvantages of the RIA included the long incubation time, lack of robustness and reliability, and poor sensitivity and precision. The potential superiority of a commercially available two-site ImmunoRadioMetric Assay (IRMA) with respect to sensitivity, precision and working range was confirmed during local evaluation but was invalidated for routine use in the screening programme by its cost. Accordingly, this project was initiated in order to manufacture an in-house IRMA which would be available to NHS laboratories. Polyclonal antibodies were generated by immunizing mice intra-peritoneally with 10 microgrammes (ug) of AFP (in amniotic fluid) suspended in Freund's adjuvant. Two boosts were given at two weeks and four weeks, and a pre-fusion intravenous boost of 4 mug was administered 3 days before post-mortem removal of the spleen. Monoclonal antibody-secreting cell lines were produced by polyethylene glycol 1500 fusion of mouse spleen cells 1:1 with mouse myeloma X63 Ag8. 653 cells. Non-hybrid myeloma cells were killed by inclusion of hypoxanthine-aminopterin-thymidine (HAT) in the medium and non-hybrid spleen cells were not viable in the culture conditions. Preparatory work on the fusion technique had shown that the presence of mouse peritoneal macrophages increased the likelihood of finding viable hybrid colonies from 6 per 96 wells to >50 per 96 wells, and rapidly reduced the quantity of debris present. Thus cells from each fusion experiment were plated into 480 wells previously inoculated with 3,000 macrophages per well. In order to investigate only high avidity/affinity antibodies, an RIA method was used to screen supernatant from all wells for antibody secretion. All supernatants demonstrating high binding were then assessed at an appropriate dilution for displacement. For this the fluid was incubated in the presence and absence of antigen. Supernatants showing considerable displacement were selected for further investigation; the cell lines were cloned by limiting dilution and large quantities of monoclonal antibody produced by injecting the clone into mouse peritoneum. Of nine mice immunized, one was killed before it was realised that the myeloma cells were non-viable owing to high incubator temperature. Two mice were used for six fusion experiments, and over 90% of the 2880 wells plated showed viable, hybrid colonies. 1136 wells (39.4%) contained anti-AFP secreting colonies (arbitrarily defined as >10% binding of iodinated-AFP), many of which will have contained several clones. Supernatant from 291 wells (10.1%) bound greater than 30% of added label, and were subjected to displacement testing. 67 wells were selected for storage after some growth, of which 16 wells, exhibiting the best growth-rate, antibody avidity and secretion characteristics, were cloned by limiting dilution. Two episodes of infection, one by Pseudomonas (from a bought-in reagent) and one by an unidentified bacterium, prevented more than two clones being used to generate large quantities of monoclonal antibody in mouse ascitic fluid for solid-phase coupling and iodination studies. The two antibodies were of similar avidity and had similar specificity, which prevented the development of a double-monoclonal assay. Characterization of the antibodies showed that they were of the IgGl isotype and had avidities of approximately 10 IRMA assay conditions were optimized, starting from a straightforward, low-cost base. A large-scale evaluation of the completed technique was performed using both density gradient sedimentation, and saline wash methods for separating the bound and free fractions. Development of a practicable IRMA protocol showed that incubation time of 2.5 hours gave minimum drift, and that the addition of sheep serum and detergent to the incubate was necessary to minimise non-specific binding effects and maximise sensitivity. (Abstract shortened by ProQuest.)