Abstract
The many physiologic roles of human alpha-fetoprotein (HAFP) and its correlation with perinatal distress/pregnancy outcome are rarely addressed together in the biomedical literature, even though HAFP has long been used as a biomarker for fetal birth defects. Although the well being of the fetus can be monitored by the measurement of gestational age–dependent HAFP in biologic fluid levels (serum, amniotic fluid, urine, and vaginal fluids) throughout pregnancy, the majority of clinical reports reflect largely second trimester and (less likely) first trimester testing due to regulatory clinical restrictions. However, reports of third-trimester and pregnancy term measurement of HAFP levels performed in clinical research and/or investigational settings have gradually increased over the years and have expanded our base knowledge of AFP-associated pregnancy disorders during these stages. The different structural forms of HAFP (isoforms, epitopes, molecular variants, etc.) detected in the various biologic fluid compartments have been limited by antibody recognition of specific epitopic sites developed by the kit manufacturers based on antibody specificity, sensitivity, and precision. Concomitantly, the advances in elucidating the various biologic actions of AFP are opening new vistas toward understanding the physiologic roles of AFP during pregnancy. The present review surveys HAFP as a biomarker for fetal distress during the perinatal period in view of its structural and functional properties. An attempt is then made to relate the AFP fluid levels to adverse pregnancy complications and outcomes. Hence, the present review was divided into two major sections: (I) AFP structure and function considerations and (II) the relationship of AFP levels to the distressed fetus during the third trimester and at term.
Introduction
Human alpha-fetoprotein (HAFP) is a tumor-associated fetal glycoprotein involved with both ontogenic and oncogenic growth (1, 2). The fetal protein is a 69-kDa single-polypeptide chain that contains 3%–5% carbohydrate and is produced in the yolk sac and fetal liver. It exhibits a triplicate domain structure configured by intramolecular loops dictated by disulfide bridging, resulting in a helical V-or U-shaped form observed in electron dot maps (3). Mammalian AFP has been classified as a member of a three-domain, cysteine-rich translated protein of the albuminoid gene family that currently consists of four members: albumin (ALB), vitamin D–binding protein, AFP, and α-ALB (4, 5). In the clinical laboratory, HAFP has long been employed both as a postoperational tumor monitoring agent and as a gestational age–dependent fetal defect marker demonstrating utility in screening for neural tube defects and aneuploidies (Table 1; Refs. 6–8). However, the main biologic role of AFP during pregnancy remains controversial to this day.
Presently, a vast biomedical literature has amassed concerning the use of HAFP during pregnancy as a biomarker in human maternal serum (MS) and amniotic fluid. Such studies have addressed the measurement of serum levels of AFP outside the normal levels in the sera of pregnant women; such values are indicative of multiple congenital malformations of the embryo and fetus. The first developmental abnormalities to be associated with abnormal AFP levels were neural tube defects and brain/spinal cord malformations (9, 10). Later, other types of birth defects were found to reflect discordant AFP levels, including chromosomal abnormalities (anaploidies) and various anatomic congenital disorders (11, 12). It was subsequently determined that assaying additional analytes together with HAFP increased the prenatal screening detection rates (Table 2). While MS-AFP levels associated with neural tube fetal defects are elevated, the chromosomal disorders demonstrate low serum AFP concentrations. Following the association of increased AFP levels with neural tube defects, additional structural anomalies have been classified within the AFP-elevated level category (6, 13). AFP serum levels during pregnancy also have been used as an ancillary aid in the diagnosis of pregnancy-related hematologic disorders (anemias), placental abnormalities, fetal death, growth restriction/retardation, and preterm labor (14). However, there exists a paucity of reports that have attempted to correlate such pregnancy anomalies with the multiple biologic activities attributed to AFP in the last decade.
Objectives.
The focus of the present report has addressed the increasing number of reports associated with abnormal AFP serum and amniotic fluid levels concom-mittant with perinatal complications and adverse pregnancy outcomes. This review has attempted to link the many functional roles of AFP with physicochemical stress/shock conditions of the fetus during the perinatal period of pregnancy. Moreover, the various physiologic roles attributed to AFP in the regulation of growth and differentiation during fetal/neonatal development have not kept pace with the increase of clinical research reports. Thus, the objectives of the present review are two-fold. First, the biologic activities of AFP during late pregnancy will be addressed, since many reviews have focused solely on AFP as a second-trimester fetal defect biomarker. Second, the various properties of AFP in light of its investigational utility as a clinical biomarker during perinatal development will be surveyed. Thus, the multitude of fetal malformations, congenital anomalies, and genetic diseases with adverse pregnancy outcomes associated with AFP in late pregnancy will be addressed. For more intense accounts of the physicochemistry and genetics of AFP, the reader is directed to earlier reviews on AFP (15–19). Recently, a report on the functional mapping of the various domains and motifs of the HAFP molecule has been put forward (20). In view of the above discussion, the present review has been divided into two major sections. These will include the following: (I) structural and functional aspects of AFP and (II) the relevance of AFP fluid levels as a biomarker in the perinatal period to adverse pregnancy outcome.
I. Structural and Functional Aspects of AFP
Structural Variants.
Molecular variants of mammalian AFP have been reported in the scientific literature since the 1970s. Some of these earlier variant forms were attributed to carbohydrate microheterogeneity and isoforms associated with varying isoelectric points (21, 22). Later reports demonstrated AFP forms that were genetic isoforms and lectin glycoforms demonstrable by electrophoretic and chromatographic procedures (23, 24). Still other variants were detected following high-pressure liquid chromatography using lectin, heavy metal, and hydrophobic solid-phase separation methodology (25). The advent of monoclonal antibodies permitted the detection and analyses of epitopic domains and subdomains that comprise the overall antigenic determinant sites on AFP (26, 27). Finally, the discovery and characterization of the molten globule forms of AFP have provided a new level of understanding regarding the various folding transition forms of this fetal protein (28).
Molecular variants of HAFP have further been reported as a result of clinical assays that detected aberrant molecular forms. Several such reports of aberrant AFP molecules first appeared in the clinical cystic fibrosis literature, resulting in confusion of the clinical usefulness of AFP for this genetic disorder. In the 1970s and 1980s, prior to the development of monoclonal antibodies, polyclonal antibody assays were not as precise and sensitive as today’s immunoassays. Such factors resulted in disparate baseline levels of HAFP in the sera of nondisease, healthy adult patients, which ranged in concentrations from 5 to 20 ng/ml. In addition, a previously reported cationic form of HAFP has been confirmed to be HAFP complexed with immunoglobulin M molecules; this cationic form has now been described in several independent laboratories (1, 25, 29, 30). Aberrant forms of HAFP also have been detected in the reproductive and urinary tracts in various clinical patients, as well as in the sera of human patients (breast cancer, reproductive disorders, etc). A nonsecreted form of HAFP lacking the N-terminal signal sequence segment was recently reported in recombinant AFP studies employing yolk sac tumors (31). Truncated forms of HAFP (~50,000 Daltons) have further been detected in cell cultures comprising hepatomas, testicular embryonal carcinomas, and breast tumors (30). Variant forms of HAFP transcripts from nontranslated regions of the AFP mRNA have recently been reported in CD34+ hematopoietic progenitor cells derived from mesodermal germ cells (32). These latter investigations described two variant forms of HAFP mRNA that are not expressed in mature cells. The variant AFP mRNAs differed from the authentic transcripts by incorporating exons from the 5′-untranslated region of the HAFP gene. The abnormal AFP transcript was found only in bone marrow, thymus, and brain tissue. The various folding intermediate forms of HAFP have recently been investigated using bacterial and yeast recombinant methodology (33, 34). The folding of both glycosylated (yeast) and nonglycosylated (Escherichia coli) forms of recombinant HAFP was studied following protein purification from aggregation-prone inclusion bodies. After AFP was denatured, it readily refolded under dilution, redox, reactions, and enzyme-linked immunoabsorbent assay (ELISA) conditions in both of the recombinant-produced AFP forms. In summary, the denaturation of recombinant-derived HAFP was found to be a reversible process independent of its starting source, fatty acid relationship, and glycosylated state.
Biologic Roles.
Determination of the biologic roles of mammalian AFP has been a research objective for many years. Similar to albumin, serum AFP is known to bind and transport a multitude of ligands, such as bilirubin, fatty acids, retinoids, steroids, heavy metals, dyes, flavonoids, phytoestrogens, dioxin, and various drugs (35, 36). Indeed, AFP has been shown to bind in vitro many substances, some of which serve as ligands for members of the steroid/thyroid nuclear receptor superfamily (37–39). However, other ligands that bind to AFP (rodent and human) include metabolic stains, L-tryptophan, warfarin, triazine dyes, phenylbutazone, streptomycin, phenytoin, anilinonaphthaline sulfate, heavy metals, low–carbon-chain alcohols, and polyunsaturated fatty acids (1, 25). Although the physico-chemical and structural properties of HAFP have been extensively described, it was mostly the in vitro functional roles that have been confirmed to date. Thus, the physiologic properties of HAFP have encompassed mainly ligand carrier/transport functions, but modulation of the immune response assays has been widely addressed (see below). Interestingly, it is growth regulation that has recently emerged as an important function of human AFPs as well as in other mammals (see below).
During the last decade, a multitude of studies have established AFP as a regulator of ontogenic and oncogenic growth (1, 25). In fact, it is the growth-modulating activity that distinguishes AFP from albumin, a major blood protein carrier/transport molecule of the albuminoid gene family. Reports now support the concept that native, full-length AFP is largely a growth-enhancing protein whose overall activity is enacted through a cyclic AMP–protein kinase A activation pathway (40, 41). However, growth is a process that requires fine tuning for both upregulation and down-regulation to operate correctly over defined time periods, such as pregnancy. Although sustained growth of the fetus is required for full-term pregnancy, the fetus does encounter situations that require periods of temporary or prolonged growth cessation, such as differentiation, transformation, and the prevention of organ/tissue overgrowth (42). Furthermore, the fetus may experience pulses of stress/ shock insults in the microenvironment compartments of both the extracellular and intracellular fetal mileux. Thus, fetal growth in a tissue or the extracellular matrix may require a temporary halt until fetal homeostasis is achieved and/or until compensated signal transduction pathways are reestablished via adaptor/scaffold protein–protein interactions. Such stress/shock encounters involving AFP include environmental extremes of osmolality, pH, oxygen tension, ischemia, glucose shock, osmotic pressure, anemia, anoxia, and excessively high ligand (steroids, fatty acids, etc) concentrations (25).
The growth regulatory properties of AFP have aroused investigational interest in studies of ontogeneic and oncogenic growth in both cell cultures and animal models. A multitude of reports have now documented that HAFP is capable of regulating growth in reproductive, hematopoietic, placental, hepatic, inflammatory, and lymphatic cells (25, 43–46). Since the late 1990s, AFP is viewed as a protein associated with modulating cell proliferation, differentiation, regeneration, and transformation in both ontogenetic and oncogenic growth processes (1, 6). Although an AFP gene knockout in mice resulted in infertile female offspring and not lethality or developmental arrest (47), a similar outcome may not necessarily be true for human beings. Although such an investigation cannot be ethically pursued in humans, clinical cases of AFP congenital deficiency have been reported in the literature (13, 48). Patients in such studies were asymptomatic and presented with normal development in their clinical histories.
In its native form, HAFP displays largely growth-enhancing properties, regardless of whether the tissue is of fetal or postnatal origin; it is the ligand-free form of HAFP at physiologic dose levels that has been shown to enhance tumor growth (49, 50). HAFP has further been shown to possess proangiogenic properties that promote neovascularization and growth in both fetal and tumor tissues (51, 52). Recent findings further indicate that HAFP can also stimulate the expression of certain oncogenes (c-Fos, c-Jun, and n-Ras) which, in turn, enhances the proliferation of human carcinoma cells (53). Finally, HAFP has been shown to functionally impair dendritic cells, inducing immune dysfunction and apoptosis of antigen-processing cells (54). In the latter report, the authors suggested a mechanism by which hepatoma cells could escape immunologic surveillance as a result of cells bearing AFP molecules on their cell surfaces.
The immunoregulatory functions of HAFP have long been studied (55). In brief, full-length HAFP has been found to be immunosuppressive in both B- and T-cell lectin blast cell stimulation (56, 57). However, recent studies have reported that not all self versus nonself AFP-specific T-cell clones are deleted during ontogeny, and that potential AFP antigenic sites persist and are recognized by both murine and human T cells. During the last several years, multiple research groups have succeeded in mapping T-cell immunodominant epitope sites on HAFP (58–62). These research groups have determined that four major human leukocyte antigen A epitotopic sites and several more minor epitopic determinants can be localized throughout the three-domain structure of HAFP.
The proposed concept that normal human pregnancy is actually a controlled state of inflammation has recently been validated in the biomedical literature (63). The human conceptus has classically been viewed as a foreign (nonself) object in the mother’s body and has long been considered a tissue allograft residing in the maternal uterus. Investigators have recently shown that the conceptus resides in an immunologically privileged site situated in juxtaposition to the placental cells (barrier), which are in direct contact with uterine tissue containing natural killer cells of the maternal innate immune system. In turn, the maternal natural killer cells secrete cytokines that attract maternal lymphocytes to the placental boundary, causing the maternal tissues to view the foreign cell clusters as tissue inflammatory sites. Such lymphocytes wall off intruder cells from the maternal tissues at the placental/uterine interface, and the cells of the conceptus are viewed by maternal cells as sites of tissue inflammation. Obviously, any fetal protein that contains cell surface proteins resembling the many cytokines of the immune system would have a definitive advantage in sustaining and maintaining the foreign tissue as a site of inflammation. AFP is such a protein consisting of a series of successive modular cassette-like peptides serving as immune system cytokine mimics, substitutes, and backup immunoregulatory peptides (1). Since the first reports emerged in the 1970s, AFP has long been recognized as both a B- and T-cell immunoregulatory protein (64).
AFP Influence on Genital Function.
Mammalian AFP has been shown to affect postnatal and adult genital function, including the onset of puberty, menstrual cycling, and spermatogenesis. AFP concentrations are low in prepubertal mammals, but supplemental injections of purified rodent AFP have been shown to inhibit follicular maturation and ovulation in the ovary and spermatogenesis in the testis (65). Rodent AFP was found to reduce the number of gonocytes during fetal and postnatal life rather than stopping oocyte meiosis at the diplatene stage, as previously proposed (66). During pregnancy, follicular maturation could be blocked by administration of purified rodent AFP at the antral stage; the follicles at this stage contain degenerating oocytes that are AFP positive following immunofluorescent staining (67). AFP was first proposed and then confirmed to play a blocking role in genital cycling as well as the induction of follicular atreas (68). Prior to that report, AFP had been localized both in the ovary and hypophysis of prepubertal rats (69). During postnatal rat development, it is known that AFP serum levels drop to adult (low) concentrations coincident with achievement of ovarian maturity on Postnatal Day 35, and administration of supplemental AFP alters reproductive events at this time (70). Finally, studies now suggest that high fetal AFP levels during human pregnancy may contribute to congenital disorders, which can include cryptorchidism in term pregnancies (71).
Studies in rodents have further demonstrated that neonatal AFP prevents circulating estrogens from accessing the brain, thus defending it from masculinization and defeminization (72–74). Such events occur in the first week of life in rodents and in the late third trimester of humans. Rodent AFP, which binds E2 with high affinity, has been proposed to either prevent entry of the estrogens into the brain or to actively transport estrogens in the developing female brain (74). Using AFP gene knockout mice, investigators have demonstrated that these mice displayed both defeminization and masculization traits. Injections of the aromatose inhibitor 1,4,6-androstene-3,17-dione into the rat brain rescued the animals from defeminization, indicating that AFP may serve to protect the rodent female brain from such effects (75). These same investigators had earlier demonstrated that female AFP knockout mice were sterile at birth due to underexpressed genes of the gonadotrophin-releasing peptide hormone, resulting in downregulation of the pituitary hormone pathways (76). Since HAFP binds little or no E2 and rodent AFP is a high-affinity estrogen binder, it is difficult to reconcile the relevance of these results to the human state. However, if the basis of estrogen brain protection is based on cell regulatory signal transduction cascades rather than E2 binding to AFP, then steroid ligand binding is not the crucial event in these studies.
AFP and AFP Receptor Expression in the Normal Human Placenta.
The authors of several studies have investigated the role of the placenta in the transport and possible synthesis of ATP in situ (77–80), since an earlier report showed that AFP was synthesized by the first-trimester human placenta (81). The objective of these investigations was to probe for the placental expression of AFP and its receptor in trophoblast cells of normal pregnancies at full term. Lafuste et al. had previously (82) reported that AFP was synthesized in the first-trimester placenta but not the term placenta; nonetheless, the presence of AFP could be histochemically localized in the villous tissue of the term placenta, presumably in the course of transplacental passage. Also, they found no evidence of mRNA AFP in the placental tissues or MS in term pregnancies. Since they also detected the presence of the AFP receptor in the placental villous tissue, the authors proposed a receptor-mediated transport mechanism for AFP placental transfer to maternal tissues. Similar to albumin, AFP transplacental passage might involve a temperature-sensitive process that does not depend on an intact cell cytoskeleton but is associated with a megalin/clathrin-mediated receptor endocytosis pathway localized to the villous trophoblast cells (83). Thus, the presence of the AFP receptor in the term placenta signified a means of transplacental passage of AFP to the maternal decidual tissues. Knowledge of the presence of AFP in the placenta is deemed important, since fetomaternal hemorrhages can occur during chorionic villus sampling in the first trimester (84).
II. Adverse Pregnancy Complications/Outcomes Related to HAFP Levels
Even though AFP was thought to be the “gold standard” biomarker for neural tube defects, elevated AFP levels had been used since 1976 as an indicator for additional perinatal distress conditions, such as bilateral renal agensis, fetomaternal transfusion, preeclampsia, intra-uterine growth retardation (IUGR), and fetal demise (Table 3; Refs. 85–87). In many instances, AFP accumulates in a biologic compartment (such as amniotic fluid) by: (i) leakage from fetal serum and cerebrospinal fluid (NTD); (ii) exposure of blood vessels in extruding viscera, leading to transudation of AFP (exopthalmos); (iii) expedited protein filtration and passage into urea (congenital nephrosis); (iv) impaired fetal swallowing or digestion in amniotic fluid (GI anomaly); and (v) altered or obstructed transplacental passage, such as in placenta accreta (Table 3). The early developmental malformations reported in the literature were structural in nature, and late-pregnancy complications were directly life-threatening to the fetus and, oftentimes, the mother. Such conditions included severe preeclampsia, premature labor, intrauterine and/or perinatal death, preterm birth, fetal wastage, and trophoblast abnormalities, including placental previa and disruption (88–92). Nonpathologic elevations of AFP in pregnancy can be the result of physiologic or procedural phenomena such as twining or multiple pregnancy, low birth weight, prematurity, or incorrect gestational age dating (93–96). Fetal defects and malformations can also be parsed by classifying them according to high or low levels of AFP in biologic fluids (Table 3). Elevated serum and amniotic fluid AF–AFP levels are usually indicative of the presence of an anatomic lesion, such as those observed in NTD, anencephaly, ventral wall defects, gastrointestinal atresia, renal anomalies, polyhyramnios and oligohyramnios, cystic hygromas with fetal hydrops, teratomas, blastomas, and disruption of placental barriers (97–103).
At the opposite extreme, low AFP levels signify the presence of chromosomal abnormalities (aneuploides), such as trisomies, as well as fetal loss, hydatidiform mole, hydrocephalus, diaphragmatic hernias, Turner’s syndrome, choroid plexus cyst, duodermal atresia, renal pyelectasis, and fetal growth restriction (104–108). Again, the fetal malformation and adverse conditions can be classified according to the calculated multiple of a population AFP median compared with the patient’s AFP median value. Low-AFP disorders categorized by various investigators have included low birth weight, multiple pregnancies, fetal wastage, perinatal death, spontaneous abortions, stillbirths, and neonatal deaths (109–111). An association was also found between second-trimester low HAFP levels and subsequent Sudden Infant Death Syndrome (SIDS). The investigators in this latter study suggested that the risk of SIDS might be mediated in part through impaired fetal growth and occurrence of adverse preterm birth events (112). Finally, the measurement of MS-AFP together with the introduction of computer-assisted Doppler measurement, which is indicative of absent or altered diastolic arterial flow in critical tissue areas, has been a major advancement in prenatal monitoring technologies.
The application of Doppler velocimetry as an adjunct to perinatal screening programs has recently increased in clinical usage. Elevated MS-AFP has now been correlated with reduced uteroplacental blood flow observed in the uterine artery (113). A study for second-trimester screening for preeclampsia was recently reported which showed that higher MS-AFP levels of human chorionic gonadotrophin (hCG), Inhibin-A, Activin-A, and AFP were accompanied by increased rates of the Doppler prediastolic notch and the derived uterine artery resistance index (114). Even though a test sensitivity of 70%–93% and a specificity of 87%–98% was achieved, the addition of Doppler velocimetry only slightly improved the predictive efficiency of the total biomarkers when used alone. The measurement of fetal middle-cerebral artery Doppler velocity has also been employed, together with MS-AFP and fetal hemoglobin as biomarkers to predict the risk of fetal anemia (115). Investigators have indeed found significant correlations between MS-AFP and both Doppler arterial measurement (r = 0.56) and fetal hemoglobin (r = 0.71) levels. In cases of alloimmunized pregnancies with fetal anemia, measured MS-AFP elevations preceded the presence of increased Doppler velocity by nearly 3 weeks. In contrast to the previous report, another group reported that both first- and second-trimester biochemical markers of trisomies had no relationship to maternal hemoglobin concentrations; however, Doppler velocity measurements were not done in this instance (116). Finally, the application of Doppler methodology to the analysis of factors predicting antepartum stillbirth was studied in conjunction with MS-AFP and pregnancy-associated plasma protein-A (PAPP-A) levels (117). Antepartum stillbirth, the single most common cause of perinatal death, has been previously associated with fetal abnormalities, congenital infections, Rh-isoimmunization, and pregnancy complications, such as IUGR, preeclampsia, and placental abruption (118). Both MS-AFP and PAPP-A measurement used in conjunction with Doppler indices of resistance to flow were found predictive of antepartum stillbirth. Both MS-AFP and PAPP-A levels are involved in placental passage and functional dynamics, because the risk of stillbirth in late pregnancy may be a result of a placental dysfunction of the placenta in early pregnancy. The invasion of the trophoblast into the uterine vessels is associated with decreased resistance to flow in the uterus, and impaired placentation is reflected in high-resistance Doppler flow velocity waveforms recorded from the uteroplacental circulation (119).
The increased usage of uterine artery Doppler (UAD) and placental ultrasound has aided in the elucidation of unexplained discordant levels of MS-AFP in the perinatal period of pregnancy. However, in the third trimester, the root cause of an AFP-associated pregnancy disorder can often be attributed to some form of placental dysfunction. Pregnancy complications related to placental disease include: (i) preeclampsia, (ii) intrauterine growth restriction, (iii) placental abruption, (iv) fetal death, and (v) spontaneous preterm labor/birth (94). Both AFP and hCG alone or in combination with other analytes can increase these associations and their subsequent risks (Tables 2 and 4; Refs. 120, 121). Premature labor and subsequent preterm delivery constitute some of the major contributors of perinatal death in the world (119). Adverse perinatal complications, such as fetal death or preterm delivery, are often attributable to chronic uteroplacental vascular insufficiency and placental infarction. Elevated MS-AFP levels are frequently found to be associated with reduced uteroplacental blood flow detected by UAD measurements (114). Moreover, abnormalities such as placental shape and/or texture in conjunction with elevated MS-AFP have been correlated with poor pregnancy outcome.
In a recent Canadian study, combined elevations of AFP and hCG were employed to predict severe placental complications using UAD, ultrasound, and placental morphology at the 19- to 23-week period (Table 4; Ref. 120). Relative risk data from this combined analyte study revealed the presence of abnormalities in studies using both placental ultrasound and UAD correlated with multiple perinatal complications, such as preterm delivery, IUGR, intrauterine fetal death, and preeclamptic pathology. Their clinical results showed a 10-fold increase in abnormal UAD scans, underscored by a high rate of underlying chronic placental vascular pathology that limited the maternal blood supply and damaged the nutrient-exchanging placental villi. These investigators further reported instances of cases displaying severe preelcampsia in addition to patients exhibiting the HELP (Hemolysis, Elevated liver enzymes, and Low Platelets) syndrome. The authors stated that their test may be able to identify the majority of women destined to either lose the fetus or deliver a preterm baby due to uteroplacental vascular insufficiency.
Other investigators have attempted to correlate abnormal second-trimester MS-AFP levels with adverse pregnancy outcomes. One group determined that abnormally high MS-AFP levels were associated with low birth weight, prematurity, and antepartum hemorrhage, whereas abnormally low unexplained MS-AFP correlated with macrosomia (overgrowth disorder) and advanced gestational age at delivery (111). While the negative predictive value of this test was high (96%), the positive predictive value was disappointingly low (9%–12%). Thus, the use of this assay to formulate a treatment plan was not promising; however, the test might find value in reassuring women about their pregnancy outcome. Another group of investigators tested whether several amniotic fluid protein components (insulin-like growth factor 1 [IGF-1], IGFβP-1, leptin, AFP) correlated with the severity of IUGR (122). Among the candidate proteins, only an association with AFP levels was found to be significant. Thus, only elevated AFP levels in amniotic fluid were useful in the early (14–18 weeks) detection of populations at risk for developing IUGR. The Quad test (AFP, hCG, UE3, and Inhibin-A) was also analyzed as a predictor of adverse pregnancy outcome in association with preterm birth, IUGR, preeclampsia, and fetal loss (123). Although it was determined that the use of multiple markers had a relatively low sensitivity and positive predictive value, it was superior to using an individual screening marker alone. Finally, the use of high MS-AFP and low amniotic fluid–AFP in the second trimester (15 weeks) indicated that diagnostic ultrasound imaging should be applied in the third trimester. Subsequent sonography at 33 weeks in this latter study revealed the presence of polyhydramios, echogenic amniotic fluid images, and gastric dilatation without gross anatomic malformations. The female karotype in this study was normal, as were the amniotic fluid Ache and AFP levels. However, at a term of 35 weeks, the newborn had pyloric atresia and displayed cutaneous blisters and erosions confirmed as epidermolysis bulbosa (124), which has previously been associated with elevated MS-AFP levels (125).
The combination of prolactin (PRL) and human placental lactagen (HPL) measurements together with an AFP determination has also been employed in the clinical diagnosis of premature rupture of membranes (PROM; Refs. 126, 127). Although such protein determinations have been measured in MS and amniotic fluid, greater sensitivity has been achieved using vaginal secretion specimens up to 41 weeks’ gestation with the MS-AFP measurements determined by ELISA assay. Using 30 ng/ml as the cutoff level, researchers reported sensitivity and specificity at 98% and 99% confidence levels, respectively (128). The investigators of this latter report stated that the results obtained with AFP alone compared with other clinical tests and echographic (ultrasound) observations were significantly better than measurement of pH together with diamino-oxidase assays and determination of PRL levels. However, in an earlier study using vaginal fluids (VFs) compared to MS and urine, VF levels of PRL and AFP were found to be 2- to 10-fold and 5- to 50-fold higher, respectively, than in paired MS and/or AF specimens. Especially in the PROM state after the 33rd week of pregnancy, VF levels of MS-AFP were predominantly higher than those found in MS, extending up to 5500 μg/ml (129). There have also been attempts to employ AFP, HPL, and PRL biomarkers in early pregnancy, although they met with less success (130, 131). In summary, only PROM studies involving HPL and AFP have continued to display their usefulness when employed in late pregnancy or at term (132).
Concluding Statements
This review has focused on the physiologic roles of AFP and its utility as a biomarker to predict perinatal distress and adverse pregnancy outcomes. Since the discovery that AFP was tumor associated in the mid-1960s, the functional roles of AFP have slowly emerged concomitantly with its ever-growing use as a biomarker in the clinical laboratory. Even though the quantitative serum levels of AFP do not always correlate with increasing size of endodermally derived tumors, the use of AFP as a tumor marker has not abated, even to the present day. Its popularity as a fetal birth defect marker increased dramatically in the 1970s and 1980s and achieved prominence in the screening of neural tube defect and chromosomal anomalies. Although AFP is not employed as a biomarker for Down syndrome in the first trimester, its association with Trisomy-21 in the second trimester provided the underpinnings for the advances in other MS marker development in first-trimester Down syndrome pregnancies.
The discoveries of discordant AFP levels correlated with perinatal distress and adverse pregnancy outcomes slowly emerged in reports emanating from the late 1980s–1990s. With each passing decade, the physiologic roles of AFP had gradually increased, and only scant attempts were made to merge those functions with the multitude of congenital malformations that had been reported. Still prominent is the long association of AFP with immune function, which is now beginning to come into greater prominence, suggesting a role of AFP in maintaining the fetal/placental unit in a controlled state of inflammation. In the future, we can expect the role of AFP in maintaining the fetus as an allograft in the mother’s body to become more clear as its relationships to the cytokines and the natural killer receptors are unraveled.
I extend my thanks and gratitude to Lynda M. Jury for her commitment and her time expenditure in the excellent typing and processing of the manuscript, references, and tables of this report.
Prenatal Screening Timeline for AFP Employed as a Biomarker Alone or in Combination with Other Analytes a
Detection Rate Percent (%) Produced by Increasing the Number of Analyte Screening Components in Prenatal Assays with the False-Positive Rate Constant at 5% a
Pregnancy Stages/Conditions with Abnormal Levels (High/Low) of HAFP a
Summary of Univariate Data Analysis Showing the Prediction of Severe Placental Complications in Women with Unexplained Combined AFP and/or hCG for Term Pregnancies a
Footnotes
This work was supported by internal funds from the Wadsworth Center.