New discoveries on the biology and detection of human chorionic gonadotropin

In 1920 Hirose showed a hormonal link between a human placental hormone and progesterone production by corpus luteal cells [1]. The name human chorionic gonadotropin (hCG) was formulated for the hormone. The promotion of progesterone production by corpus luteal cells was assumed to be the principal function of this hormone. Until recent years this has been assumed to be the sole function for hCG.

The first pregnancy test, the rabbit test, was formulated [2, 3] in the 1920s. For four decades bioassays such as the rabbit test were the only practical way to measure hCG or detect pregnancy. In 1960 with the development of polyclonal antibodies came the agglutination inhibition test [4]. Then, in 1967 with discovery of the competitive immunoassays the hCG radioimmunoassay was developed [5‐8]. This became the first rapid and sensitive test and led to the dawn of commercial hCG tests as seen today. hCG testing became part of the evaluation of every pregnancy. The initial radioimmunuassays used an antibody to whole hCG α β dimer. The α-subunit of hCG is identical with the α-subunit of LH. As such the initial RIA detected both hCG and LH limiting its use for the early detection or pregnancy. In 1973 the hCG β-subunit radioimmunoassay was introduced, specifically detecting hCG through its β-subunit [9]. This led to sensitive and specific pregnancy tests, detecting pregnancy soon after missing menses. The discovery of monoclonal antibodies in 1975 was paramount to the development of modern immunometric tests [10]. Two-antibody immunometric assays for hCG arose in the nineteen eighties, and with them came sensitive antibody enzyme labeling and high sensitivity fluorimetric and chemiluminescent tracers [11‐14]. These are the formats of assays used in commercial laboratories today.

In 1970 hCG was shown to be a non-covalently linked dimer [15]. The 1970s saw the determination of amino acid sequence of hCG subunits (Figure 1), and the observation that hCG contained 4 N-linked and 4 O-linked oligosaccharides [16, 17]. The 1980s and 1990s saw the determination of the structures of the N- and O-linked oligosaccharides on hCG as produced in pregnancy and gestational trophoblastic diseases (Figure 2) [18, 19], it saw the elaboration of the hCG subunit gene structures [20], and to our understanding of the hCG/LH receptor and the mechanisms of hCG endocrinology whereby hCG promotes progesterone production [21, 22].

Was this the completion of the hCG physiology, biochemistry and immunoassay story? In the past 10 years a revolution has occurred with this molecule. Firstly, with the finding that the polypeptides that make hCG do not just make one biologically active molecule, but a group of three biologically important molecules, regular hCG, hyperglycosylated hCG and hyperglycosylated hCG free β-subunit. Each of these three molecules having different physiological functions and each having had separate roles in the evolution of humans. We have also seen in the last 10 years the elaboration of our understanding of pituitary hCG and new data on hCG and trophoblast function. Furthermore, there has been the discovery of numerous new applications when measuring only hyperglycosylated hCG and only the hyperglycosylated and regular free β-subunit. This review examines these hCG variant molecules and how the past 10 years they have changed the way they are viewed. How hCG is a key element in the evolution of the human brain, and how today we look differently at defined hCG variant assays and the specificity of total hCG assays in general.

Regular hCG predominates in all normal and abnormal pregnancies (Table 1). It is the principal hCG variant produced through the bulk period of pregnancy (Table 2), it is present at reduced levels during spontaneous aborting and ectopic pregnancy [14] and at double regular levels in Down syndrome pregnancy [23, 24]. Regular hCG is also the principal molecule produced in individuals with hydatidiform moles or pregnancies comprising solely trophoblast tissue [18, 25]. Regular hCG is also normally produced by the pituitary gland at the time of LH peaks and in menopausal women [25, 26].

Multiple variants of hCG have been detected in serum and urine samples (Figure 3). A free α-subunit of hCG and a free β-subunit of hCG (hCG free β) have been demonstrated in pregnancy serum, cancer patient serum and urine and cancer cell line culture fluids (Table 1) [27, 28]. A terminal degradation variant of hCG β-subunit is found in urine samples. This is β-subunit core fragment, it comprises two fragment of β-subunit, β6–40 and β55–92, held together by disulfide bonds (Figure 3) [29]. Urine β-core fragment is used as a general tumor marker for all non-gestational malignancies [30‐34]. A large form of hCG dimer is synthesized by choriocarcinoma cells [15, 18, 35]. This is hyperglycosylated hCG with 1.5 fold larger N-linked and double size O-linked oligosaccharides (Figure 3) [18, 19, 36]. Similar unduly large N- and O-linked oligosaccharides have been demonstrated on the hCG free β produced by non-gestational cancer cells, we call this molecule hyperglycosylated hCG free β [20, 35], Similar large oligosaccharides are found on the α-subunit secreted in pregnancy, free α-subunit [37]. Just as hCG is hyperglycosylated in choriocarcinoma and gets larger oligosaccharide side chains, free α-subunit is further glycosylated in the malignancy, gaining an additional O-linked oligosaccharide, we call this molecule O-glycosylated free α-subunit [38].

The regular hCG and hyperglycosylated hCG degradation pathways involve elastase and other proteases secreted by macrophages associated with tumor tissue or present in the circulation [39]. These cleave or nick hCG at β-subunit 44–45 or 47–48 [18, 39], generating a nicked regular hCG and a nicked hyperglycosylated hCG, a nicked hCG free β and nicked hyperglycosylated hCG free β (Figure 3). These same enzymes progress further to cleave and release the C-terminal peptide on the nicked molecules (βCTP, β residues 93–145) generating nicked hCG missing βCTP, nicked hyperglycosylated hCG missing βCTP, nicked hCG free β missing βCTP and nicked hyperglycosylated free β missing βCTP [18, 25, 39]. This is 15 variants, all found in serum and urine samples in normal or abnormal pregnancies, gestational trophoblastic diseases or non-gestational malignancies (Figure 3).

Of these 15 variants, just 5 are natural synthetic products made by the placenta in pregnancy or by non-trophoblastic malignancies, regular hCG, hyperglycosylated hCG, hyperglycosylated hCG free β and free α-subunit and O-glycosylated free α-subunit. The other 10 variants are degradational product, from macrophage cleavage and cleavage by proteases in the circulation and the kidney. Trophoblast cells are the active cells of the placenta (Figure 4). Fused trophoblast cells or syncytiotrophoblast produce the hormone regular hCG [40]. These same cells make an excess of α-subunit which is secreted as free α-subunit [37, 40]. Initial studies indicated that cytotrophoblast cells hyperglycosylated hCG [40, 41]. Hyperglycosylated hCG is an autocrine factor, it functions separate to regular hCG to promote invasion at implantation of pregnancy and malignancy in gestational trophoblastic diseases [41‐44]. Recent studied by Handschuh and colleagues show that hyperglycosylated hCG is made only by the extravillous invasive cytotrophoblast cells [44], the cell that terminate anchoring villi and invade the myometrium. These same cells make an excess of α-subunit which is secreted as free α-subunit and O-glycosylated free α-subunit [37, 40]. It is noteworthy that free α-subunit has no known biological function, it is a biological waste product [37].

Non-gestational malignancies produce hyperglycosylated free β [14, 31‐34], as demonstrated this acts as an autocrine factor directly promoting cancer cell growth and malignancy. Clearly, 3 of the 5 synthesized variants of hCG are independent molecules, hormones or autocrines, with separate functions, regular hCG, hyperglycosylated hCG and hyperglycosylated hCG free β.

Hyperglycosylated hCG is the extravillous invasion cytotrophoblast cell invasion signal [40‐44], and is produced in states characterized by cytotrophoblast cells or invasion (Table 1). It is the principal or sole form of hCG produced in the first week of pregnancy, following implantation of the fetus (Table 2) [43, 45‐47]. Unduly low levels of hyperglycosylated hCG clearly mark failing pregnancies, whether a biochemical pregnancy or early pregnancy loss, spontaneous abortion or ectopic pregnancy [43, 45‐47]. While minimally present in hydatidiform mole cases, rising proportions of hyperglycosylated hCG mark invasive hydatidiform mole (invasive mole) and development and advancement of choriocarcinoma (Table 1) [41‐44, 48‐50]. Hyperglycosylated hCG can also differentiate invasive and non-invasive (quiescent trophoblastic disease) gestational trophoblastic diseases [43, 49]. Hyperglycosylated hCG similarly marks retrodifferentiation to cytotrophoblast cells and cancer advancement in testicular germ cell malignancy cases [42, 48]. Hyperglycosylated hCG has also been used as an improved marker for fetuses with Down syndrome, in both the first and second trimesters of pregnancy [51, 52]. It is also an outstanding marker for predicting third trimester hypertensive disorders and preeclampsia in the second trimester of pregnancy [53].

It is important to fully understand the differentiation of trophoblast cells which separate production of hyperglycosylated hCG and regular hCG (Figure 4). Stem cytotrophoblast cells differentiate to make extravillous invasive cytotrophoblasts [54], which produce hyperglycosylated hCG (Figure 4). They also differentiate to make villous cytotrophoblast cells which do not produce hyperglycosylated hCG(Figure 4). These fuse to form multinucleated syncytiotrophoblast cells, with 2 to 50 nulei [54]. The fusion of cytotrophoblast cells is controlled and promoted by regular hCG [55]. Cytotrophoblast cells are the principal cells present at the time of pregnancy implantation. These cells rapidly differentiate to extravillous invasive cytotrophoblast, and villous cytotrophoblast, which fuse to syncytiotrophoblast as pregnancy advances (Figure 4) [54, 55]. In choriocarcinoma and testicular germ cell malignancies cytotrophoblast cell predominate. It appears that the carcinogenesis process le makes a non-differentiating non-villous invasive cytotrophoblast cells producing hyperglycosylated hCG [42]. Thus hyperglycosylated hCG is a potent marker for these malignancies. In Down syndrome pregnancies, trisomy 21 limits the differentiation of villopus cytotrophoblast cells leading to an accumulation of these cells and hyperglycosylated hCG [56]. This is apparently due to a failure of regular hCG-promoted differentiation [56]. It is for this reason that hyperglycosylated hCG is a marker of Down syndrome in the first and second trimester of pregnancy.

Hyperglycosylated hCG free β is produced at the time of implantation in very early pregnancy, and at very low proportions throughout the length of pregnancy (Table 2). While present at low levels, it is widely used in predicting Down syndrome in the first and second trimesters of pregnancy [57]. The hyperglycosylated hCG free β is produced in other or non-trophoblastic malignancies. A portion of almost every human malignancy advanced, retrodifferentiates and produces hyperglycosylated hCG free β [30‐38, 58‐62]. Hyperglycosylated hCG free β production also distinguishes placental site trophoblastic tumors from other forms of gestational trophoblastic disease (Table 1).

All told, regular hCG, hyperglycosylated hCG and hyperglycosylated hCG free β are widespread in pregnancy, pregnancy abnormalities, trophoblastic malignancies and other or non-trophoblastic malignancies, with each form predominating in different disorders (Table 1). These 3 molecules plus ten degradation products constitute the 13 forms of hCG β-subunit present in serum or urine samples under different conditions (Figure 3). Ideally a total hCG tests should detect all of these hCG-related molecules to optimal monitor pregnancy, gestational trophoblastic diseases and cancer cases.

The peptide structure of the hCG group of molecules was established by Bahl and colleagues in 1972 [63] and confirmed and refined by Morgan and colleagues in 1975 (Figure 1) [64]. Oligosaccharides constitute approximately 25–30% of the molecular weight of regular hCG. The N- and O-linked oligosaccharide structures of regular hCG were first determined by Kessler and colleagues [65, 66], and refined and confirmed by Mizouchi and Kobata in 1980 [67], further refined by Elliott and colleagues in 1997 [18] and by Kobata and Takeuchi in 1999 [19]. As illustrated in Figure 1, regular hCG, molecular weight 36,000, comprises a 145 amino acid β-subunit and a 92 amino acid α-subunit. There are 2 N-linked oligosaccharides on the α-subunit of hCG and 2 N-linked oligosaccharides on the β-subunit of hCG. There are also 4 O-linked oligosaccharides on the C-terminal peptide region of the β-subunit of hCG (Figure 1).

Trisaccharide O-linked oligosaccharides are present on the C-terminal peptide segment of regular hCG (Figure 2) [18, 19]. A Full Bi (biantennary) and a Partial Bi structure are present at the 2 N-linked oligosaccharide sites on the α-subunit of regular hCG and a Full Bi and Full Bi w/fuc biantennary structure are present at the 2 N-linked oligosaccharide sites on the β-subunit of regular hCG [18, 19].

Between 1985 and 1997 Cole and colleagues examined the structure of the hCG produced in invasive trophoblastic disease, choriocarcinoma, and showed it to be different to that on regular hCG [18, 68, 69]. As found, this invasive cell hCG, whether from early pregnancy or choriocarcinoma had double size hexasaccharide O-linked oligosaccharides and larger (triantennary) N-linked oligosaccharides (Figure 2). This molecule was named hyperglycosylated hCG. No difference was observed in hCG peptide sequence, only in the oligosaccharide side chains [18]. Hexasaccharides replaced trisaccharides on O-linked oligosaccharides of hyperglycosylated hCG. Full Tri (triantennary), Full Tri w/fuc and Partial Tri structures replaced the Full Bi, Full Bi w/fuc and Partial Bi biantennary oligosaccharides at the N-linkage sites on hyperglycosylated hCG. These increased the molecular weight of hyperglycosylated hCG from approximately 36,000 (regular hCG) to approximate 40,500. The molecular weights are approximate (± 1000) due to variability in sialic acid content [18, 19]. These oligosaccharide structures on regular and hyperglycosylated hCG have all now been confirmed by two other groups using alternate oligosaccharide structure methods [19, 36, 42, 70].

Variations are found in the N-linked and O-linked oligosaccharide structures on hCG with differences in cellular metabolism and different expression of glycosyltransferase activities (18,19,36). In 1997 it was demonstrated that the 4 O-linked oligosaccharides are the principal difference between choriocarcinoma or testicular germ cell malignancy hyperglycosylated hCG and pregnancy regular hCG [18, 42]. There are 2 principal types of O-linked oligosaccharides (Figure 2), trisaccharide and hexasaccharide. While first trimester normal pregnancy urine hCG contained 12.3 to 19% of the hexasaccharide sugar structure (regular hCG, n = 6 individuals, mean = 15.6%), choriocarcinoma urine hCG contains 60 to 100% hexasaccharide structures (hyperglycosylated hCG, n = 6 individuals, mean = 74.2%), 5-fold more than first trimester pregnancy [18]. A smaller difference was observed with the 4 N-linked oligosaccharide structures on pregnancy and choriocarcinoma molecules [18]. While 6 first trimester pregnancy samples contained an average of 10.3% triantennary structures at the four N-linked sites on hCG, 5 choriocarcinoma cases contained 33% of the triantennary structures (3-fold more than first trimester pregnancy) [18, 19, 42]. The largest difference between regular hCG and hyperglycosylated hCG is at the O-linked oligosaccharides.

In 2006 the oligosaccharides were evaluated for the first time on a site by site basis using mass spectrometry methods with multiple regular hCG and hyperglycosylated hCG preparations from pregnancy, choriocarcinoma and testicular germ cell malignancies [36]. Researchers found a greater fucose content on N-linked oligosaccharides in cancer cases [36]. They demonstrated site-specific difference in O-linked oligosaccharides. They showed the constant presence of predominantly hexasaccharide structures in pregnancy and choriocarcinoma at Serine residue 121 [68–89%], with variable structures on Ser 127, 132 and 138, primarily trisaccharide structures on regular hCG in pregnancy and hexasaccharide structures on hyperglycosylated hCG in choriocarcinoma [36]. In conclusion, combining results, it appears that differences in hCG β-subunit O-glycosylation at 3 sites, Ser 127, 132 and 138, are the principal discriminator of regular hCG and hyperglycosylated hCG [18, 19, 36].

Here we refer to pregnancy hCG as regular hCG and choriocarcinoma hCG as hyperglycosylated hCG. With the advent of hyperglycosylated hCG antibodies and very specific immunoassays [71, 72], this larger form of hCG was identified in urine samples considered too dilute in hCG for analysis of oligosaccharide structures. By these methods, hyperglycosylated molecules were shown to be 96% of total hCG in urine samples and 89% in serum samples (Table 2) at the time of embryo implantation [43, 45, 46, 73]. The proportions of hyperglycosylated hCG molecules rapidly decline in the weeks following implantation, averaging 68% and 49% in urine and serum samples at the time of missing menses, 50% and 36% at 5 weeks of gestation, 25% and 21% at 6 weeks of gestation, and < 1.5% in the second and third trimesters of pregnancy (Table 2) [45, 46, 74].

The free β-subunit of hCG detected in choriocarcinoma patient urine and in urine from individuals with non-trophoblastic neoplasm is larger than the β-subunit isolated from regular hCG dimer. Larger oligosaccharide side chains have been indicated [75‐77]. Examination of the structure of the free β-subunit of hCG reveals molecules with a high proportion of fucosylated trianntenary N-linked oligosaccharide structures and hexasaccharide O-linked oligosaccharides, analogous to hyperglycosylated hCG [36]. This explains the larger molecular size for hCG free β observed in cancer cases. The name hyperglycosylated hCG free β is adopted for this molecule. It appears that the free α-subunit of hCG receives larger oligosaccharide side chains than the α-subunit of hCG dimer [37], and that the free β-subunit like the free α also is hyperglycosylated.

In the previous section we discussed the structures of the regular hCG, hyperglycosylated hCG, hyperglycosylated hCG free β and free α-subunit made by trophoblast and cancer cells. Here we discuss the dissociation, cleavage and clearance of these hCG forms, or the synthesis and structure of the hCG degradation intermediates. Figure 5 summarizes the dissociation, cleavage and circulatory clearance pathways of hCG, hyperglycosylated hCG and hyperglycosylated hCG free β [39, 78‐88].

As an initial step, the dimers regular hCG and hyperglycosylated hCG are either slowly cleaved (nicked) or slowly dissociated into α and β subunits (Figure 5). The dissociation half-time of dimers into subunits is 700 ± 78 hours at 37°C [78]. Hyperglycosylated hCG dissociates much more rapidly that regular hCG (140 ± hours at 37°C [39]).

A macrophage or circulating leukocyte elastase protease cleaves hCG β-subunit at Val residue 44 (residue 44–45 cleavage) or Gly residue 47 (residue 47–48 cleavage) generating nicked hCG (Figure 1) [39, 78]. Nicked hCG is 30-fold less stable than regular hCG (dissociation half time 22 ± 5.2 hr. at 37°C [78]) rapidly dissociating releasing a free α-subunit and a nicked hCG free β. Similarly, dissociated hCG β-subunit is rapidly cleaved by leukocyte elastase to make nicked hCG free β. Both nicking and dissociation eliminate hCG hormone activity (shown by ability to bind corpus luteum hCG/LH receptor and to promote progesterone production [39]). Dimer is cleared from the circulation very slowly (clearance half-life 35.6 ± 8.0 hr. [82, 84]), hCG free β is cleared 10-fold faster (3.93 ± 0.68 hr. [83, 84]). As such the combination of nicking and dissociation, in either order, leads to the rapid deactivation and clearance of regular or hyperglycosylated hCG [78].

Nicked hCG or nicked hCG free β, while cleaved at 44–45 or 47–48, remain structurally intact due to 5 disulfide linkages between the component peptides. Further degradation of hCG by leukocyte elastase leads to cleavage of β-subunit at Leu residue 92 [39]. The peptide 93–145 the βCTP is not held by disulfide linkages and leads to nicked dimer missing the βCTP or nicked hCG free β missing βCTP [18, 32, 39]. Nicked dimer missing the βCTP is rapidly degraded to nicked hCG free β missing the βCTP (Figure 5). The nicked hCG free β missing the βCTP made by either pathway is rapidly degraded in the kidney by exoproteinases and glycosidases to β-subunit core fragment, the terminal degradation product of regular hCG, hCG free β, hyperglycosylated hCG and hyperglycosylated hCG free β [85‐88]. β-subunit core fragment is β-subunit residues 6–40 linked to residues 55–92, held together by 5 disulfide linkages. It has degraded N-linked oligosaccharides, missing the non-reducing terminal sialic acid, galactose and N-acetylglucosamine residues and fucose residues so terminates in mannose [85]. As shown, injection of pure CHO cell recombinant regular hCG (contains no nicked, hyperglycosylated or dissociated hCG components) into the human circulation leads to excretion of β-subunit core fragment in the urine [88], clearly confirming the degradation pathway.

Nicking and cleavage to make β-subunit core fragment continues throughout pregnancy deactivating hCG [4]. In first trimester pregnancy serum, 9% of hCG dimer molecules are nicked, all free β molecules are also nicked, in the last 2 months of gestation 21% of dimer molecules are nicked and all free β molecules are nicked [39, 78]. In pregnancy urine samples, β-subunit core fragment concentration or the degradation end product concentration is relatively small in the first weeks of gestation, it equals hCG dimer concentration at 6–7 weeks of pregnancy. β-core fragment concentrations then exceeds dimer concentrations in urine thereafter, 7 weeks to term [14, 89, 90]. Mean β-core fragment concentrations average 58% of urine hCG concentrations (mean proportion) at 5 weeks gestation, 105% at 7 weeks gestation rising to 305% of hCG concentrations in the final month of pregnancy [14, 89, 90]. β-subunit core fragment is the principal hCG related molecule in pregnancy urine samples. In choriocarcinoma cases, 100% hyperglycosylated hCG may be present in serum samples, yet either the hyperglycosylated hCG or β-subunit core fragment may predominate in parallel urine samples [91, 92].

All non-trophoblastic malignancies produce primarily hyperglycosylated hCG free β [31, 30‐34]. In most cases, as a result of macrophage elastase activity at tumor site, the degradation product β-subunit core fragment is detected in corresponding urine samples [14, 93‐95]. Considering the extremely low levels of hyperglycosylated hCG free β produced by non-trophoblastic malignancies and their rapid clearance from the circulation, urine β-subunit core fragment may be a more sensitive or more measurable tumor marker than serum hyperglycosylated hCG free β [14, 93‐95].

It is common to find residual nicked hCG or nicked hCG missing βCTP in serum or β-subunit core fragment in urine, the products of degradation, in the weeks following parturition of pregnancy or surgical evacuation of hydatidiform mole or tumor. It is important in these cases to detect nicked hCG and nicked hCG missing βCTP in monitoring serum and β-core fragment in monitoring urine following evacuation of ectopic pregnancy, parturition, evacuation of hydatidiform mole and choriocarcinoma and non-trophoblastic malignancy malignancy to be sure that all tissue is removed.

In the preceding section we examine the structures of regular hCG, hyperglycosylated hCG and hyperglycosylated hCG free β as produced by cells. In this section we show how elastase and other proteases and glycosidase convert 3 forms of hCG β-subunit in serum and urine into 13 forms (Figure 3). Different secreted and degraded forms of hCG may best mark different conditions (Table 1). We emphasize again, it is important when considering a total hCG assay and the structure and degradation findings presented here, to select an assay detecting all form of hCG β-subunit. Regular hCG may mark a normal pregnancy, hyperglycosylated hCG may best mark gestational trophoblastic diseases, hyperglycosylated hCG free β may best mark a Down syndrome pregnancy, nicked hCG missing βCTP may best mark clearing hCG at parturition or following termination, and urine β-subunit core fragment ay best mark non-trophoblastic malignancies.

Chorionic gonadotropin (CG) is part of a family of hormones that includes LH, follicle stimulating hormone and thyroid stimulating hormone. These are the glycoprotein hormones that share a common α-subunit coded by a single gene, and a separate β-subunit which dictates hormone function [130, 131]. In 1980 Fiddes and Goodman [130], examined the DNA sequence for the β-subunits of CG and LH and showed that the evolution of CG from LH was by a single deletion mutation in LH β-subunit DNA and read-through into the 3'-untranslated region occurring in early simian primates. In 2002 Maston and Ruvolo [131], examined the DNA sequences of the β-subunit of CG in 14 primates and showed that the genes to make CG and its variants were not present in prosimians or primitive primates (example: Lemur), but evolved by the indicated deletion mutation with early simian primates (example: platyrrhine or new world monkey). Multiple mutation leading to DNA and amino acid sequence changes then occurred with the continued evolution of the CG β-subunit genes from early simian primates to the advanced simian primates (example: pongo or orangutan) and further changes with the evolution of hominids [131]. In summary, prosimian LH has 3 N-linked oligosaccharides mean pI 8.4, from the deletion mutation CG evolved in early simians with 5 oligosaccharides mean pI 6.3, from a further mutation advanced simians evolved a CG with 6 oligosaccharides mean pI 4.8, and with further mutations hominids evolved with 8 oligosaccharides mean pI 3.5 (Table 3).

The acidity of the evolving CG (and its variants) with additional oligosaccharides very much affects its metabolic clearance rate or circulating half times and thus serum concentration and its effective bio-potency [132, 133]. As an example, at one extreme, regular human CG has 8 O-linked and N-linked oligosaccharides all terminating in sialic acid residues. These acidify hCG resulting in a molecule with a mean isoelectric point (pI) of 3.5, and a circulating half time of 35.6 hours [82, 84]. At the other extreme, asialo hCG with no acidic sugars on oligosaccharides has a pI of 8.5 and a circulating half time of 5.7 minutes [133]. Because of the acidity due to the sialic acidic residues on each N- or O-linked oligosaccharides, regular hCG circulates for approximately 400 times longer than asialo hCG, raising the circulation concentration proportionately.

The earliest CG form in early simian primates had 2 O-linked oligosaccharides on the β-subunit at serine residues 121 and 132 and one N-linked oligosaccharide at residue 30 [131]. This yielded a molecule with a mean pI of 6.25 [131, 135]. As a result of a point mutation at residue 127 (Asn-Ser), the advanced simian primate evolved with a CG having 3 O-linked oligosaccharides [131, 134, 135], and a mean isoelectric point of 4.9 (Table 1) [135]. With the evolution of hominids, a point mutation occurred at residue 138 and 15 and CG molecules were developed with 4 O-linked and 4 N-linked oligosaccharides [131]. The 4 O-linked 4 N-linked oligosaccharide molecule, is the most acidic CG, pI 3.5 [82]. We deduced from isoelectric points that the 2 and 3 O-linked oligosaccharide 3 N-linked oligosaccharide CGs produced by early and advanced simian primates had minimal and middling circulating half times within the wide half time range of LH at 25 minutes and human CG at 35.6 hours. It is concluded that with evolution from early simians to advanced simians to humans, and the stepwise increases in isoelectric point that circulating half-times and blood concentrations of CG and hyperglycosylated CG increased logarithmically.

Brain size in mammals is directly related to the combination of body mass and the metabolic support of the developing progeny [136]. The greater brain size, seen in advanced primates and hominids, correlates with disproportionately large energy demands by the developing fetuses [136‐142]. Numerous studies support the concept that advanced primates, and to a greater extent humans, have had to develop ultra efficient placentation mechanisms to support the increasing nutritional demands of the embryonic brain (Table 3) [136‐146].

As shown in Table 3, the prosimian primate had an average size mammalian brain, 0.07% of body mass. In this species, epitheliochorial placentation was sufficient. This involves the placenta loosely attaching to the wall of the endometrial stroma with no invasion. Nutrients and oxygen had to diffuse through multiple myometrial and endometrial stromal layers to transfer from the maternal to fetal circulation. Hemochorial placentation started with the early simian primate with the evolution of CG [137, 139, 140, 146]. In all likelihood regular CG and hyperglycosylated CG initiated and drove hemochorial placentation [119]. The earlier simian primate had 2 O-linked oligosaccharides and 3 N-linked oligosaccharides, was not very acidic, so probably cleared the circulation quickly leaving extremely low concentrations of circulating molecules to promote invasion and angiogenesis, leading to inefficient hemochorial placentation. The placenta invaded through the thickness of the decidua only. Placentation supported nutritional transfer necessary for a brain of 0.17% body mass or 2.5 fold greater than prosimian primates (Table 3) [137, 139, 140, 146]. The advances simian primate had 3 O-linked oligosaccharides, was more acidic, so cleared the circulation at a mediocre rate leaving mediocre concentration of circulating molecule to promote invasion and angiogenesis. This permitted invasion through the decidua to one tenth the thickness of the myometrium [137, 139, 140, 146]. Placentation supported nutritional transfer necessary for a brain of 0.74% body mass or 4.5 fold greater than the early simian brain (Table 3) [137, 139, 140, 146].

Placentation and nutrition transfer was taken to the extreme in the hominids. Hominid CG is very acidicwith 4 O-linked and 4 N-linked oligosaccharides with a circulating half-time of 35.6 hours leading to highly elevated levels of regular CG and hyperglycosylated CG. This led to invasion through the decidua to one third the thickness of the myometrium and to ultra efficient angiogenesis. Placentation supported nutritional transfer necessary for a brain of 2.4% body mass or 3 fold greater than that of advanced simians (Table 3) [137, 139, 140, 146].

Considering the relationship between regular CG, hyperglycosylated CG and hemochorial placentation, and between advancing acidity of CG and advancing invasion and angiogenesis, it would not be unreasonable to propose that the evolution of CG in early simians started primates on the evolution path to advanced brains, or is at the root of the evolution of humans.

The evolution of hyperglycosylated hCG free β appears to an end result result of the evolution regular and hyperglycosylated hCG. The gene for the β-subunit evolved in the human genome. The gene for the α-subunit existed separately as the α-subunit of glycoprotein hormones. When cancer cells regress and retro-differentiate they express cytotrophoblast proteins. By default hyperglycosylated hCG free β is expressed which promotes cancer growth and malignancy [30, 32, 38, 58, 61, 62, 59, 124] leading to poor prognosis [32, 34, 122, 123].

The human CG- and hyperglycosylated CG-driven placentation model may be optimal for fetal nutrition and brain development. In many respects placentation efficiency has expanded to the extreme to accommodate humans supporting a 2.4% body mass brain versus a 0.07% brain in prosimians and many other mammals. Failure of this extreme process, however, may be associated with life threatening complications. Ineffective invasion into the myometrium leads to inefficient hemochorial placentation. This leads to nutritional deficits and anoxia to the fetus compensated for by preeclampsia in the mother [138, 142, 144, 145, 147, 148]. Preeclampsia has been shown to be linked to unduly low hyperglycosylated hCG levels at the time of initiation of hemochorial placentation at the end of the first trimester [149]. Preeclampsia is one of the most life-endangering complication of pregnancy and occurs uniquely in humans with the fetal demand for deep implantation and the oxygen and nutritional mechanisms needed to support brain development.

With the implantation demands of the fetus, pregnancy failures occur more commonly in humans that in any other species. Pregnancy failures, miscarriages and biochemical pregnancies or early pregnancy losses, account for 41% of gestations in humans compared to rodents (10%) and to a similar low percentage in all other species [43, 145, 150]. Shallow implantation or ineffective invasion leads to human pregnancy failures [145]. Approximately two thirds of human failures can be attributed to inappropriate invasion [145] and are associated with unduly low hyperglycosylated hCG concentrations [45, 73, 151]. All pregnancies with normal term outcome produce a significant proportion of hyperglycosylated hCG at the time of implantation [43]. Similarly, two thirds of pregnancies that fail are associated with significant low hyperglycosylated hCG concentrations [43]. Together this data, and the established invasion promoting function of hyperglycosylated hCG suggest that the two thirds of failures due to placentation are the same as the two thirds due to low hyperglycosylated hCG. It is inferred that the high incidence of pregnancy failures, in humans, is a consequence of deficient hyperglycosylated hCG.

Hydatidiform mole does occur in primates [152, 153]. Invasive moles, gestational trophoblastic neoplasms or choriocarcinoma, however, have only been observed in humans. These are malignancies of trophoblast tissue associated with extravillous invasive cytotrophoblast cells, driven by the production of CG-H [41, 42, 146‐149]. It appears that just as hyperglycosylated hCG promotes invasion of by normal pregnancy extravillous invasive cytotrophoblast cells during implantation, it is also responsible for the malignancy-like invasion by cytotrophoblast cells in trophoblastic neoplasia. Circulating levels of hyperglycosylated CG are significantly raised by acidic glycosylation, from early simian to advanced simian primates, and then raised further by acidic glycosylation with the evolution of hominids. We deduce that only humans have circulating hyperglycosylated CG levels at magnitudes that can cause malignant-like states. It is inferred that invasive gestational trophoblastic diseases occurs uniquely in humans as a complication of the hominid generation of high concentrations of the invasion promoter hyperglycosylated CG

Considering the relationship between hyperglycosylated hCG and trophoblast invasion, and the proposed link between evolution of hyperglycosylated CG and primate and human evolution, a different perspective emerges on problems with human fecundity due to pregnancy failure, on preeclampsia in human pregnancy and on invasive gestational trophoblastic diseases. Do these findings and hypotheses open new avenues for preventing pregnancy failure and preeclampsia, and for treating or preventing invasive gestational trophoblastic disease?

Is it conceivable that pregnancy failures can be prevented by administration of human CG-H at the time of implantation. The abundance of research showing a link between hyperglycosylated hCG and pregnancy failure supports such trials [43, 73, 151]. The lack of availability of a pure or clinical usefully preparations of hyperglycosylated hCG at the present time, however, limits the initiation of such trials.

We question similarly the potential use of hyperglycosylated hCG therapy in reducing the risk of preeclampsia and gestation induced hypertension in high risk pregnancies. It is possible that administration of hyperglycosylated hCG at 10–13 weeks of gestation would ensure appropriate completion of invasion and appropriate activation of hemochorial implantation. Risks, such as those which may cause over-invasion or placenta percreta would need to be taken into account. The possible prevention of these disorders could have significant effect on pregnancy complications.

The relationship between human hyperglycosylated hCG and invasive gestational trophoblastic disease is widely proven [41, 42, 146‐149]. Therapy would need to involve the blocking of hyperglycosylated hCG. This may be very much simpler than administrating the invasion-promoting agent. Three separate studies using nude mice transfected with choriocarcinoma cells have shown that blocking hyperglycosylated hCG (choriocarcinoma hCG) production with hyperglycosylated hCG antibody or genetic methods, halts all growth and development of malignancy [41, 103, 104]. The next step is to perform analogous studies in humans, using human antibodies to hyperglycosylated hCG or vaccines to similarly block malignancy. These would initially be used in patients with chemotherapy resistant cases of choriocarcinoma, and then later examine the prevention of invasion or malignancy in patients with hydatidiform mole.

Hyperglycosylated hCG free β is produced by most human malignancies solely as a consequence of the evolution of regular and hyperglycosylated hCG for promotion of hemochorial placentation. When cancer cells regress and retro-differentiate they express cytotrophoblast proteins. By default hyperglycosylated hCG free β is expressed which promotes cancer growth and malignancy [30, 32, 33, 58, 61, 62, 59, 124] leading to poor prognosis [32, 31, 122, 123]. Efforts have been directed toward using different free β derivatives as vaccines in the treatment of non-gestational malignancies. Success has been reported, with free β immunity improving cancer outcome or cancer survival [125‐129]. The use of site specific free β vaccine technology suggests a plausible route to the development of adjuvant cancer therapies specifically targeting patients with free β producing non-gestational tumors. An hCG β-subunit derivative vaccine is commercially available (CG Therapeutics, Seattle, USA). It is being tested in clinical trials with different non-gestational neoplasms in the USA.

In 1976 Chen and colleagues [154], identified human chorionic gonadotropin (hCG) in the circulation of non-pregnant females. This was later demonstrated as being of pituitary origin [155]. Multiple studies with post-mortem human pituitary extracts showed that they all contained hCG at an average concentration of 3.0% of the LH concentration (average 0.8 μg per gland) [156, 157]. Individual pituitary glands contained varying concentrations of hCG, 0.05 to 3.3 μg/gland [156, 157]. These findings have been confirmed multiple times, with pituitary hCG production being demonstrated as part of normal reproductive physiology [43, 157‐164]. When considering that hCG and LH share a common α-subunit gene [165], and the LH β-subunit gene is buried amidst 7 hCG β-subunit genes [165], it may be considered as no biological surprise that the pituitary produces a small amount of hCG along with higher concentrations of LH. Pituitary hCG has been detected aside LH in the menstrual cycle [158‐162], and post-menopausal in response to the absence of estradiol feedback regulation [43, 162‐164].

Cole and colleagues recently examined regular hCG in 371 menstrual cycles (Table 4) [166]. The menstrual cycles averaged 28.6 ± 3.8 days (mean ± standard deviation) in length, with an LH peak detected at 14.6 ± 3.1 days following the commencement of menstrual bleeding. The LH peak averaged 210 ± 161 IU/L in magnitude in urine samples (Table 4). Regular hCG was detected (sensitivity 1 mIU/ml) around the time of the LH peak in (LH peak +/- 4 days) in 332 of the 371 menstrual cycles (89%) (Table 4). The average hCG concentration was 1.52 ± 0.91 IU/L in positive LH peak urine samples. The majority of positive regular hCG urines corresponded to the LH peak day, with lesser numbers of urines corresponding to 1 to 4 days before and 1 to 4 after the LH peak. A correlation was observed between LH concentration and the corresponding hCG concentration among the 332 urines, equating hCG and LH by linear regression, r² = 0.97; by Bartholomew's test of ordered means, p = 0.004 [166].

It was concluded that 1.0 mIU/ml or a greater concentration of regular hCG was produced parallel to the LH peak in 89% of menstrual cycles [166]. It was inferred that regular hCG was also produced in the 11% remaining menstrual cycles but at levels < 1 mIU/ml. Regular hCG promotion in the pituitary may be incidental, a byproduct of gonadotropin releasing hormone (GnRH) promotion of LH and follicle stimulating hormone (FSH) expression. Regular hCG expression, however, may be purposeful. LH has a very short circulating half life, 0.42 hrs, while hCG has a long circulating half-time 40 hrs [81, 167]. In urine regular hCG is approximately 1% of the concentration of LH. In the pituitary gland, however, regular hCG exists at 3% of the concentration LH [156, 157]. In serum, 2-fold higher levels of regular hCG are observed than recorded in urine [14], while 4-fold lower levels of LH are observed [168, 169]. As such, regular hCG levels in serum may be much more significant than observed in urine (approximately 7% of LH concentration). The serum concentration represents the concentration acting on ovarian cells. Regular hCG may supplement the promotion of ovulation by extending the LH peak range. This suggest that slow clearing regular hCG may also have a function in boosting the rise in progesterone levels during the beginning of the luteal phase of the cycle. We infer that regular hCG is a functional hormone during the menstrual cycle, either in promoting ovulation or promoting progesterone production.

The pituitary processes molecules slightly differently to syncytiotrophoblast cells, It terminates some N-linked sugar structures in N-acetylgalactosamine-sulfate rather than in galactose-sialic acid (Figure 2). The 4 N-linked oligosaccharides on regular hCG produced by the pituitary terminate in N-acetylgalactosamine-sulfate, while the 4 O-linked oligosaccharides are like those made by syncytiotrophoblast cells terminate in galactose-sialic acid [170]. Because of the difference in N-linked oligosaccharide structure and its acidity, pituitary regular hCG is half as biologically potent as syncytiotrophoblast cell regular hCG, due to a shorter circulating half-life [171].

As recorded in all medical text books, LH and FSH levels are controlled by negative feedback to the hypothalamus by estradiol, controlling GnRH pulses during menstrual periods. As also recorded, estradiol feedback becomes reduced during perimenopause years and disappears during menopause. With this reduction, LH and FSH are normally produced at much higher levels without control at these menopause times. What textbooks do not mention, is that hCG produced at the time of the LH peak also becomes more evident in serum in perimenopause and postmenopause [43, 162, 164, 166, 172‐176].

The USA hCG Reference Service is a clinical reference service specializing in unexplained hCG production and in gestational trophoblastic diseases. Despite the 20+ publications in major medical journals on the normality of positive serum hCG in perimenopause and postmenopause women and in those having had bilateral oophorectomy, physicians continue to be needlessly concerned, considering gestational trophoblastic disease and cancer as possible options. The USA hCG Reference Service has investigated 319 cases with persistent low hCG results in the absence of history gestational trophoblastic disease or other malignancy (Table 5). These included 112 woman producing hCG that was later proved to be of pituitary origin (Table 5). These were 43 woman postmenopause with amenorrhea, ages 50 to 62, with mean serum regular hCG level of 11 ± 6.2 ranging from 8 to 28 mIU/ml. There were 60 woman perimenopause with oligomenorrhea, ages 39 to 59, with mean serum regular hCG level of 9.8 ± 6.7 [172‐176] ranging from 2 to 22 mIU/ml. Finally there were 9 women of normal menstrual cycle age that had received bilateral oophorectomy for cancer-related reasons, putting them into amenorrhea, they had a mean serum regular hCG level of 10 ± 7.2 mIU/ml ranging from 7 to 25 mIU/ml [172‐176]. As shown in Table 5, of the 319 cases with persistent low hCG results, the concluding diagnoses has ranged from pituitary hCG, non-gestational malignancy, gestational trophoblastic neoplasm, quiescent gestational trophoblastic disease and false positive hCG. As shown, in the USA hCG Reference Services experience, pituitary hCG is the prime cause of persistent low hCG levels [176].

Medical centers throughout the world demand a serum total hCG test prior to any surgery, prior to many x-ray procedures and to certain drug protocols. The erroneous assumption is made by medical centers id that a total hCG test can only show pregnancy. If a woman is positive in the hCG test she is referred to an obstetrician. Unfortunately few obstetricians are aware of pituitary hCG. If a women in oligomenorrhea or amenorrhea has a positive pregnancy test prior to surgery it can mean multiple months of delays and false alarms before pituitary hCG is identified and the procedures are rescheduled. The USA hCG Reference Services has consulted on 12 cases waiting multiple years for kidney transplants. When the match was found the procedures were cancelled due to positive hCG tests. In each case multiple months passed before the USA hCG Reference Service was contacted and normal physiology pituitary hCG identified. The patients then rejoined the waiting list once more for a kidney transplant. There should be no requirement for an hCG test if a woman is over 45 years old.

Perimenopause can be difficult to diagnose. The normal symptom is oligomenorrhea. Is the hCG of pituitary origin? A recent article by Granowitz and colleagues [177] shows that the presence of an elevated FSH levels of > 30 mIU/ml justifies the presence of peri- or postmenopause and the presence of pituitary hCG. This is and evaluable method for determining whether pituitary hCG is possible. The pituitary as the source of hCG can be confirmed by placing subjects on a high estrogen oral contraceptive pill for 3 weeks. It will suppress pituitary hCG and the hCG will no longer be detectable [174‐176].

The discovery of monoclonal antibodies in 1975 was paramount to the development of modern immunometric tests [9]. Modern two-antibody immunometric assays for hCG arose in the early nineteen eighties; with them came the concepts of antibody enzyme labeling and high sensitivity fluorimetric and chemiluminescent detection assays [10‐14]. These mechanisms are the basis of the automated tests used in all commercial laboratories today. The principal of the immunometric test is one antibody (called the capture antibody) binding one site on hCG and immobilizing it and a second antibody labeled with enzyme tracer (called a tracer antibody) binding a distant site on hCG labeling the immobilize complex. The immobilized and labeled capture antibody-hcg-tracer antibody complex can then be quantified, with the amount of tracer or tracer enzyme products being directly proportional to the amount of complex or concentration of hCG. The dual antibody immunometric technologies are also the basis of all modern physician's office point of care (POC) rapid pregnancy tests and home use or over the counter (OTC) rapid pregnancy tests with one antibody immobilized in the result window on the nitrocellulose device and one antibody mixing with the serum or urines and labeled with a blue or red dye as tracer. A positive result is indicated by a line formed by the immobilized antibody-hCG-dye antibody complex.

As discussed throughout this article there are multiple variants of hCG in serum and urine in different pregnancy and cancer conditions (Table 1) [14, 27‐34, 41‐43, 48], these may be invariably detected by different types and different brands of hCG test. By definition there are today 2 clear classes of hCG immunoassay. Intact hCG assays which may or may not detect all forms of dimer (hyperglycosylated hCG, nicked hCG and regular hCG), and total hCG tests which may or may not detect hCG dimer and its free β-subunit (hCG free β, hyperglycosylated hCG free β, nicked free β, free β-subunit missing the βCTP and β-subunit core fragment). There are also three clear types of hCG tests. There are professional laboratory quantitative serum hCG tests (PRL tests), point of care or physicians office qualitative serum and urine hCG tests (POC tests), and home or over the counter qualitative hCG tests (OTC tests).

Most PRL tests used today are total hCG assays. These most commonly involve an antibody to the β-subunit core structure (tertiary structure formed by β1–92) as tracer antibody combined with an antibody to the βCTP (β92–145) as capture antibody. This insures minimal recognition of LH, which lacks a βCTP. There are 2 major limitations with assays using this antibody combination. Antibodies to the natural βCTP commonly are dependent upon the O-linked oligosaccharide structure on the natural βCTP, since this accounts for a major portion of the molecular weight of this segment. Most assays using this antibody configuration poorly detect hyperglycosylated hCG, so critical for use in early pregnancy detection and in choriocarcinoma management (Table 1), because the immunoglobulin used is against the regular hCG βCTP. While this combination usually detects nicked hCG it does not detect hCG or free β missing the C-terminal segment. Some manufacturers have replaced the antibody to the natural βCTP with an antibody to a βCTP synthetic peptide, β131–145, with no oligosaccharide component. An assay with using an antibody to the synthetic βCTP does not distinguish regular hCG and hyperglycosylated hCG. An alternative configuration used by one manufacturer involves 2 antibodies to different regions of the β-subunit core structure. This rare combination permits detection all 13 dimer and β-subunit variants (Figure 3).

We ask why doesn't every manufacturer try and optimize their hCG assay to detect the 13 variants of hCG? Why do some manufacturers sell a test that detects primarily regular hCG and poorly detect variants, yet claim there test to be an optimal hCG test. While others sell tests that detect the13 of 13 dimer and β-subunit variants claim nothing more? As will become apparent from reading this article, detection of 13 of 13 dimer and β-subunit variants is critical for using hCG as a pregnancy test, gestational trophoblastic disease test and cancer test (Table 1)? Unfortunately all variants of hCG are simply considered as general hCG to physicians, they order hCG tests for pregnancy, gestational trophoblastic disease and cancer cases regardless of the specificity of the test used by the laboratory. Laboratories usually order tests based on their throughput, speed and cost. The problem arises with the FDA, who regulate all hCG tests. Their guidelines are entirely based on rather dated 1960s, 1970s and 1980s information, under which hCG is one molecule and detection of only regular hCG is required. There is no requirement to detect hyperglycosylated hCG, even though we now know today that it is the principal or sole variant of hCG present in serum and urine in early pregnancy [40, 42, 43, 45, 46, 79, 178, 179], and no requirement to detect free β-subunit or its degradation derivatives (Figure 3), which we now know is the only form produced by cancers. Modern hCG test do state that the test is only approved for pregnancy detection and is not approved for managing gestational trophoblastic diseases or cancer applications (clinicians are not informed of this, it is solely in a note included in the test ingredient box delivered to the laboratory), even though the use of these application continue at every clinical laboratory, regardless of this limitation and regardless of which hCG test their laboratory uses. In fact, no test is approved by the FDA for gestational trophoblastic disease, Down syndrome screening, detecting failing pregnancies or cancer testing. In many ways gestational trophoblastic disease and cancer patients are left in the cold, with no manufacturer interest in producing an optimal test. The end result of the outdated FDA rules is that manufacturers have no interest and no financial benefit in updating or improving the specificity of their assays. In our experience the only tests made that is appropriate for all applications in the Siemen's Immulite series (Immulite 1000, Immulite 2000 and Immulite 2500, same antibody mix but different platforms designed for different throughput).

We examined the major hCG tests sold in the USA, those accounting for > 1% of the USA market (Table 6). They are all automated total hCG immunometric assays. Each assay is sold in semi-automated formats for low throughput centers, and high throughput fully automated formats for intensive use laboratories. All tests are run from a computerized platform. As shown in Table 6, all assays detect regular hCG and all may or may not detect other hCG variants. The Roche Elecsys series poorly detects anything other than regular hCG and hCG free β, and the Siemens Dimension (used to be Dade Dimension), like the Roche assay, poorly detects hyperglycosylated molecules, so critical to pregnancy. Multiple tests, such as the Baxter Stratus, Siemens AC180, Siemens Centaur and Beckman Access and DXI all inappropriately detect hCG β-subunit (not detected on equimolar basis with regular hCG) and hyperglycosylated hCG free β. The majority of tests do not detect hCG or its free β-subunit missing the C-terminal segment [79, 179] (Table 6). Only one assay, that using the 2 antibodies to the β-subunit core structure, detects all pertinent standards. As such, only one assay is useful for all pregnancy, gestational trophoblastic disease and cancer application, this is the Siemen's Immulite series (used to be DPC Immulite series). It should be noted that the author of this review, Laurence A. Cole PhD has no financial, consulting or other conflicts of interest with DPC or Siemens. The finding that the Siemen's assay is the most appropriate is strictly the result of blind unbiased testing at multiple laboratories.

As shown in Table 7 most assays, with the exception of the Siemen's Immulite series give widely varying results when used for gestational trophoblastic disease and cancer management. Looking, for instance, at a case of serous ovarian cancer in Table 7 (15^th cases from top), results varied from 1 mIU/ml (Baxter Stratus) to 115 mIU/ml (Siemen's Immulite series) dependent on the assay being used. In this case the patient was primarily producing hyperglycosylated hCG free β missing the C-terminal segment which was only appropriately detected by the Siemen's Immulite series assays. Examining a partial mole case (20^th case from top), results varied from 9 mIU/ml (Baxter Stratus), to 15 mIU/ml (Beckman Access), to 31 mIU/ml (Siemens ACS180), to 48 mIU/ml (Abbott Axsym), and 235 mIU/ml (Siemen's Immulite series) depending on the assay being used. hCG missing the C-terminal segment was also produced in this case. As shown in Table 7, almost every gestational trophoblastic disease or cancer sample is measured with extreme variability. There is no guide given by laboratories to physicians to indicate which test is used or its limitations. Physicans just assume that the test used at the medical center is good.

Similar specificity problems are seen with POC and OTC assays (Table 8). Most POC tests are for some strange reason intact hCG tests detecting dimer only. Due to the configuration of the antibodies in these devices, they poorly detect hyperglycosylated hCG (Table 8) [48]. Similarly, with one exception, manufacturers of OTC tests make devices which poorly detect hyperglycosylated hCG (Table 8). The one exception is the Church and Dwight First Response and Answer devices that equally detect regular hCG, hyperglycosylated hCG and free β-subunit [180]. It is astonishing that other manufacturers of POC and OTC devices market devices that poorly detect hyperglycosylated hCG since measurement of this molecule is critical to early pregnancy detection. How can these be called pregnancy tests? Once more the blame is placed on the FDA and their very out of date guidelines and the absence of a requirement to detect anything other than regular hCG. They basically permit all manufacturers, regardless of what they produce, to label their devices similarly "greater than 99% accurate." This leads to much misinterpretation and confusion.

We sadly conclude that the field of hCG assays is full of confusion. This is in large part due to inappropriate FDA regulations. It is important for test and platform purchasers to be weary of these limitations and to choose a test with consideration of all possible applications. It seems that only with laboratories exclusively choosing broad specificity tests like the Siemen's Immulite (a small part of the hCG market in the USA) and the Church and Dwight's First Response or Answer (OTC tests, better than all available POC tests), that the market will change and manufacturer's will be economically forced to produce better quality products. Some centers, like the Charing Cross Hospital, Gestational Trophoblastic Disease Center, in London, UK have little trust in any automated hCG test for their gestational trophoblastic disease and cancer applications [181]. They only trust results from their "in house" hCG β-subunit competitive radioimmunoassay. At the USA hCG Reference Service we will only use the Siemens Immulite. We are referred pregnancy and cancer patients with erroneous hCG results. All too often we are referred cases that have received needless therapy due to faulty hCG results, or missed therapy due to missed recurrence of disease. The specificity problem lingers on year after year, it must be addressed, and can probably only be resolved by the FDA.

Numerous publication in the past 10 years describe new applications for total hCG tests and for new assays measuring hyperglycosylated hCG alone and hyperglycosylated hCG free β alone.

The field of hCG has changed dramatically in the past 10 years. This is the focus of this review. It has moved from hCG a hormone produced by trophoblast cells that promotes progesterone production by luteal cells to hCG as group of molecules each with different functions. That is regular hCG a hormone produced by syncytiotrophoblast cells which promotes myometrial spiral artery angiogenesis, aiding nutrient supply to the placenta. Hyperglycosylated hCG is an autocrine factor made by extrauterine invasive cytotrophoblast cells in pregnancy and malignant cell non-villous cytotrophoblast cell is choriocarcinoma and testicular germ cell malignancies. This form of hCG controls placental invasion and implantation and cancer cell malignancy. Hyperglycosylated hCG free β is an invasion promoter made by non-trophoblastic cancer cells which enhances cancer cell growth and malignancy. All of these new findings have been confirmed by multiple separate laboratories. In many ways, the new hCG as we know it today, is a very unique molecule, one α and one β polypeptide structure and 3 independent bioactive molecules. This is a unique biological combination, 3 molecules differing in carbohydrate structure.

Major changes have also occurred in the past 10 years in the hCG immunoassay field, New perspectives on hCG assays and what they should detect. Multiple new application for hyperglycosylated hCG in Down syndrome screening, detecting pregnancy, detecting pregnancy failures and in the diagnosis and management of gestational trophoblastic diseases. New applications for free β in predicting Down syndrome pregnancies, in diagnoses of PSTT and as a tumor marker for all non-gestational malignancies, and new vaccine treatments blocking free β enhancing survival of malignancies.

In many respects the new discoveries are like the rediscovery of hCG as a group of molecules has occurred, a group of molecules critical to human evolution and essential to human pregnancy. With all the new doors opening in the hCG story, everything is still far from complete. Yes, we know that hyperglycosylated hCG is an autocrine factor that controls extravillous invasive cytotrophoblast growth and invasion but we are far from sure what receptor it binds. We know that its mechanism involves blocking apoptosis, but how does it promote metalloproteinases to advance invasion? Multiple indications and preliminary studies suggest a TGFβ-like receptor but this is far from proven. This needs to be resolved. Similarly, we know that hyperglycosylated hCG free β also seemingly binds a cytokine receptor and works through a mechanism involving antagonism of apoptosis, but what receptor does this bind? This also needs to be resolved. The idea evolves of using hyperglycosylated hCG to prevent failure of pregnancy and to prevent preeclampsia. It sounds incredible to consider preventing these complications of pregnancy, but at least ten years of testing in advanced primate and human clinical trials is needed before this can become realistic. Clinical trials are already ongoing using in hCG β-subunit vaccine in the treatment of malignancies. A lot of promise is in store. At the current time we do not have a standard for hyperglycosylated hCG or commercial or recombinant method for producing hyperglycosylated hCG, maybe this is a first step. A lot of improvement is needed in the coming years in hCG assay development, in improving total hCG assays to measure regular hCG, hyperglycosylated hCG and hyperglycosylated hCG free β on an equi-molar basis, and in the development of new hyperglycosylated hCG only assays, using them in IVF pregnancy testing and applying them to pregnancy failures and other promising applications. Yes, the past 10 years has shown great advances in the field of hCG, what the next 10 years may show could be even more exciting.