MONOCLONAL-ANTIBODIES TO THE FREE BETA-SUBUNIT OF HUMAN CHORIONIC-GONADOTROPIN DEFINE 3 DISTINCT ANTIGENIC DOMAINS AND DISTINGUISH BETWEEN INTACT AND NICKED MOLECULES

The present study was designed to characterize monoclonal antibodies (mAbs) specific for the free beta-subunit of hCG (free hCGbeta), to develop two-site immunoradiometric assays (m-IRMAs) specific for free hCGbeta, and to study the reactivities of various molecular forms of hCGbeta in these assays. We attempted first to delineate the antigenic regions present specifically on the free hCGbeta by studying the binding pattern of seven mAbs directed preferentially to hCGbeta, designated FBT11, P8E, P10F, HB2, P5D, P5H, and INN-64. Competitive inhibition experiments performed by RIA demonstrated the specificity of these mAbs for the free hCGbeta as noncross-reacting with the beta-subunit of human (h) LHbeta. m-IRMAs were used to analyze the arrangement of epitopes on hCGbeta. Experiments performed with the seven mAbs used either as capture antibodies or radiolabeled indicators confirmed the specificity of the seven mAbs for the free hCGbeta, and that mAbs FBT11, P8E, and P10F bound to equine LHbeta (eLHbeta), but did not bind to a fragment of hCGbeta called the beta-core fragment (betaCF). These antibodies defined an antigenic domain identified as A. In contrast, mAb HB2 bound neither to eLHbeta (ebeta) nor to betaCF and was directed to domain B (ebeta negative, betaCF negative). Finally, mAbs P5D, P5H, and INN-64 bound to betaCF, but did not bind to eLHbeta, and defined a third domain identified as C (ebeta negative, betaCF positive). Collectively, these results demonstrate that at least six antigenic domains are present on the free hCGbeta and that a limited set of amino acids was shared among these domains; domains A, B, and C are present only on the free beta-subunit, while three other domains recognized by mAbs FB19, FBT10, and 518B7 are present on both free hCGbeta and hCG. Thus, most of the surface of the hCGbeta appears to be antigenic and accessible to antibody binding. Three different m-IRMAs specific for free hCGbeta were then constructed using either mAb FBT11 (domain A) or HB2 (domain B). The study of the reactivities of various molecular forms of hCGbeta in these assays demonstrated that the recognition of hCGbeta forms nicked at position 43 (beta43), 44, and/or 51 (beta44/51) varied between assays. Indeed, the cross-reactivities of beta43 and beta44/51 in the free hCGbeta assay based on FBT11 and FBT10 mAbs (m-IRMA 11-10) were 80% and 40%, respectively, whereas these nicked forms displayed a cross-reactivity of 16% and less than 2%, respectively, in the assay based on mAbs HB2 and FBT11 (m-IRMA 2-11). m-IRMA 2-11 was used in parallel with m-IRMA 11-10 to measure free hCGbeta levels in the sera of pregnant women and patients with either hydatidiform mole or gestational choriocarcinoma. A dramatic decrease in free hCGbeta levels was observed with m-IRMA 2-11 compared to those observed with m-IRMA 11-10. These results should be taken into account when using free hCGbeta serum levels to establish a differential diagnosis between normal pregnancy and benign or malignant gestational trophoblastic disease and emphasize the necessity of the delineation of the fine specificity of mAbs to free hCGbeta used for the design of immunoassays.