| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Center for Health and the Environment,3 University of California, Davis, California 95616-8615
Institute for Biomedicine,4 Anhui Medical Institute, Heifei, Anhui 230032, People's Republic of China
Department of Environmental Health,5 Harvard School of Public Health, Boston, Massachusetts 02138
ABSTRACT
A need exists for broadly applicable biomarkers of pregnancy outcome in population-based studies that assess environmental hazards to human reproduction. Previous studies have demonstrated that during the periimplantation period, measures of the circulating levels of immunoreactive hCG (IhCG) are not predictive of pregnancy outcome, whereas measurements of the circulating levels of bioactive hCG (BhCG) provide information relating to pregnancy outcome and might provide the basis for an early biomarker of pregnancy outcome. However, for this biomarker to have broad application in population-based studies, it must be adapted to urinary hCG metabolites. The principle objective of the present study was to characterize the periimplantation excretion patterns of urinary hCG metabolites of pregnancies that resulted in live birth (LB), early pregnancy loss (EPL), and recognized clinical abortion (CAB) with an immunoenzymometric assay specific to intact hCG and an LH/chorionic gonadotropin cellular bioassay as the basis for a preliminary comparison between successful (LB) and failing (EPL and CAB) outcome groups. Automated immunoassays for FSH and hCG were used to define each conceptive cycle's implantation window. The timing of first hCG detection was significantly later for the EPL group. Pregnancies that resulted in LB had consistently rising average daily IhCG and BhCG levels, with no significant differences when average daily IhCG and BhCG measurements were compared (Student t-test, P > 0.05), whereas pregnancies that resulted in CAB and EFL had lower average daily IhCG and BhCG levels that increased inconsistently. These findings demonstrate that critical information related to pregnancy outcome may be present when multiple urinary hCG isoforms are measured. Further data suggest that the rate of change for the ratio of daily BhCG over IhCG levels might be useful as the basis of a broadly applicable early biomarker for pregnancy outcome.
human chorionic gonadotropin, implantation, placenta, pregnancy, syncytiotrophoblast
INTRODUCTION
Public concern exists that environmental chemicals may have adverse effects on human reproduction [1]. Roughly 70% of all detectable human conceptions result in live birth (LB), approximately 20% in early pregnancy loss (EPL), and 10% in recognized clinical abortion (CAB) [2]. Most EPLs likely result from chromosomal abnormalities [3], and a significant proportion may have an environmental origin [4, 5]. Sensitive immunoassays for urinary metabolites of hCG found in self-collected, daily urine samples have been used to assess the rates of pregnancy and pregnancy loss of nonclinical subjects in field studies [2, 6–9]. However, their use has been limited by gaps in our understanding of hCG secretion and subsequent excretion profiles and their relationship to the hCG in vivo function as a tropin. A more complete understanding of urinary hCG secretion patterns during the periimplantation period of normal and failing pregnancies is essential for the development of improved analytical methods and algorithms for accurately assessing pregnancy and pregnancy outcome.
A practical early biomarker for pregnancy outcome could be used by investigators in field studies to address directly the difficulty of assessing the hazard posed by a toxicant that caused a modest increase in the rate of CAB against the relatively high rate of spontaneous pregnancy loss in humans. In one scenario, pregnancies in the control and exposed cohorts of a population-based toxicology study would be identified with an hCG immunoassay. The EPL would be removed from further analysis, whereas CAB would be analyzed with an early biomarker for pregnancy outcome. Those pregnancies that resulted in CAB and that had biomarker profiles consistent with previously validated profiles for CAB and EPL would be classified as "abnormal shortly after implantation" and excluded from subsequent analysis. Those pregnancies that resulted in CAB and that had biomarker profiles consistent with previously validated profiles for LB would be identified as "normal shortly after implantation." Only the pregnancies classified as "normal shortly after implantation" but subsequently resulting in a CAB would be used to determine the relative risk of pregnancy loss for the control and exposed cohorts, because an environmental stressor likely induced them. In this way, the background 30% rate of spontaneous pregnancy loss could be removed effectively, thus allowing a subtle difference between the cohorts to have greater statistical power.
Human chorionic gonadotropin is produced by the developing trophoblast as early as 6 days after fertilization and is carried by maternal circulation to the ovary, where it can bind to the LH/chorionic gonadotropin (CG) receptors and prevent luteolysis, thereby providing hormonal support to the endometrium, which in turn provides a substrate for the developing embryo [10]. Human chorionic gonadotropin recently has been demonstrated to regulate directly the formation of placental syncytium [11].
Human chorionic gonadotropin is a heterodimeric, glycoprotein hormone consisting of noncovalently linked CGA and CGB subunits and several carbohydrate side chains that account for 30% of its total molecular weight [12]. Substantial metabolic processing of the intact hCG molecule occurs before its excretion. The predominant polypeptide isoforms in circulation (hCG, CGB, CGB-core fragment, and nicked hCG) all are present in the daily urine samples of early pregnancies [13]. Recent studies have shown that intact hCG and CGB predominate until the fifth week after conception, when CGB-core fragment overtakes them [14]. Immunoassays for hCG that are currently employed to detect EPL in population-based studies measure the amount of intact hCG alone or in combination with CGB, CGB-core fragment, and nicked hCG present via structural epitopes recognized by specific monoclonal and/or polyclonal antibodies [15].
Presently, limited correlation is observed between the measurements of hCG levels made with widely used immunoassays and the ability of hCG to bind to the LH/CG receptor and transduce signal. An LH/CG radioreceptor assay and a luminescence LH/CG cellular bioassay were developed to measure hCG molecules in pregnancy serum and to investigate the interaction of hCG with its receptor and downstream signal complex [16]. Because of the LH/CG receptor's strict binding requirement for intact molecules, hCG molecules capable of triggering the downstream signal cascade in LH/CG cellular bioassay were subsequently found to be a subset of the intact molecules measured by an immunoenzymometric assay (hCG IEMA) specific for intact hCG molecules (IhCG) and an LH/CG radioreceptor assay [17]. The LH/CG cellular bioassay and the hCG IEMA were then used to analyze five consecutive daily periimplantation blood samples collected from 14 conceptive cycles resulting in LB, seven conceptive cycles resulting in CAB, and four conceptive cycles resulting in EPL [18]. A steep rise in the concentration of bioactive hCG (BhCG) was observed across the daily samples of pregnancies that resulted in LB. On the other hand, pregnancies that resulted in CAB and EPL were observed to have deficient hCG production, as reflected in their daily IhCG and BhCG levels and reduced rates of increase for hCG concentrations across their 5-day analysis windows. Furthermore, some pregnancies that resulted in EPL and CAB produced a form of hCG with lower biological activity than hCG produced by pregnancies that resulted in LB, as reflected the arithmetic ratio of daily BhCG measurements over IhCG measurements (hCG B:I ratio) for the daily blood samples being analyzed [18]. These data were consistent with the observation that an early form of hCG with attenuated bioactivity was produced and then gradually replaced with a form of hCG having full bioactivity during the periimplantation period of pregnancy. We then hypothesized that the extent and rate at which the early form of hCG is replaced by the later form likely is the basis for the differences observed between surviving (LB) and failing (EPL and CAB) pregnancies and may be useful as the basis for an early biomarker of pregnancy outcome in population-based studies.
Based on the assumptions that immunoreactivity is a measure of the quantity of hCG glycoprotein and that bioactivity provides information relating to the ability of a measured amount of circulating hCG to function as a tropin, the daily hCG B:I ratio should provide a self-contained assessment of the average quality of individual hCG molecules in a biological sample. The main advantage of this approach is that the hCG B:I ratio is independent of the time course of the appearance and/or the increase of hCG levels in circulation, because the quality of hCG is a characteristic of individual molecules rather than of the timing or number of molecules secreted. In fact, the greatest differences originally observed between pregnancies that resulted in LB and those that resulted in CAB and EPL were reflected by the rates at which the daily hCG B:I ratios increased characteristically for each outcome group [18]. We therefore hypothesized that the rate at which the hCG B:I ratios increase during the periimplantation period may have utility as a reliable means of predicting an individual pregnancy's outcome in population-based studies.
Population-based studies that have investigated the adverse effects of environmental hazards on human reproduction with hormonal biomarkers have relied on self-collected urine samples. The endocrine patterns associated with monitoring EPL require daily samples, often over multiple menstrual cycles. Daily urine samples are easily collected, handled, and stored, and they offer feasibility where daily blood samples do not. For any potential biomarker to have real value as a tool for population-based studies, it must be adapted to the metabolites in urine samples.
In this communication, we report the results of a retrospective characterization of the levels of total and BhCG metabolites in the daily, periimplantation urine samples of surviving and failing pregnancies as a first step toward assessing the feasibility of developing an early biomarker for pregnancy outcome to use in population-based studies of reproductive toxicology. Specifically, the use of a sensitive chemiluminescent immunoassay (CIA) for urinary FSH and hCG metabolites (FSH and hCG CIA) with an automated assay platform to describe the timing of ovulation and implantation in conceptive cycles with surviving and failing outcomes was demonstrated. The above information was then used to define a physiologic implantation window for each conceptive cycle, which was necessary for the subsequent characterization of the levels of total and BhCG metabolites with an hCG IEMA and an LH/CG cellular bioassay, respectively. In this way, a conceptual approach to analyses with the hCG B:I ratio was described with laboratory methods that are presently available. This conceptual approach is potentially adaptable and scalable for future field studies, which will be necessary to validate and apply superior biomarkers based on the one described in the present study.
MATERIALS AND METHODS
Original Subjects and Sample Collection
The samples analyzed in the present communication were selected from a large prospective study of reproductive health originally conducted from 1996 to 1998 in the textile mills of Anhui, China. Study protocols were approved by the Institutional Review Board of the Harvard School of Public Health and all collaborating institutions within China. Additionally, written informed consent was obtained from each participating woman and her husband. A detailed description of the study population and the sample data collection methods has been reported previously [8]. Briefly, beginning on the day that contraceptive use was stopped and an attempt was made to conceive, each woman recorded the occurrences of sexual intercourse, vaginal bleeding, medications taken, and medical conditions in a daily diary for 12 mo or until a pregnancy was clinically confirmed. During this period, daily first-morning urine samples were collected, held temporarily at 4°C, and stored at –20°C until analysis. Urinary hCG and creatinine (CR) were then analyzed in each woman's daily urine samples with the combo-hCG immunoradiometric (IRMA) and the colorimetric CR assay as described previously [8]. Diary data, pregnancy outcome, and daily hCG data indexed by CR were included on plots for each subject.
For the analyses described in the present communication, 32 conceptive cycles that resulted in LB and five conceptive cycles that resulted in CAB were selected from 373 and 49 total cycles that resulted in LB and CAB, respectively. Cycle selection was random, but the ultimate determination of whether a cycle was included in the present study was based on the availability of sufficient sample to complete the proposed analyses. Additionally, 14 cycles were screened for EPL, and six cycles were identified and carried forward. The first day of menstrual bleeding on each selected subject's plot was used to establish the beginning of each conceptive cycle, and the combo-hCG IRMA data were used to identify cycles with positive hCG levels and no clinically recognized pregnancy for potential EPL. The approximate cycle day at which hCG was detectable by hCG IRMA was noted.
Definitions: Pregnancy Outcomes
Both LB and CAB outcomes were identified before analysis with clinical information listed on each subject's plotted hCG IRMA data. The CAB outcomes were defined as pregnancy losses that occurred after the clinical recognition of pregnancy and before Gestational Week 28. The EPL outcomes were defined as conceptive cycles with no reported clinical pregnancy and with positive hCG CIA results.
Day of Ovulation (FSH Day 0)
To establish the timing of ovulation for staging and alignment in subsequent analyses, the day and the level of the midcycle CR-indexed FSH peak were determined for each conceptive cycle with the FSH CIA [19]. The day of the FSH peak was defined as the day with the highest CR-indexed or unindexed FSH level after Cycle Day 7. The first 7 days of each cycle were excluded, because they were most likely to have transient high FSH levels, corresponding to the beginning of the follicular phase. Group average, SEM, and coefficient of variation (CV) were calculated for the peak cycle day and the corresponding FSH level with CR-indexing.
hCG CIA to Identify EPL
Following analyses with the FSH CIA, 14 cycles with no clinically recognized pregnancy and at least one daily combo-hCG IRMA level greater than 0.5 ng/mg of CR were identified and subsequently analyzed with the hCG CIA assay. Six of 14 cycles had positive hCG CIA results, as defined by hCG levels that exceeded a sliding 5-day average baseline plus 2 SD for two of three consecutive days beginning with FSH Days +1 to +5.
Definition: Day of hCG Detection
To establish the timing of implantation and the connection of embryonic and maternal circulation in the previously identified conceptive cycles, the first day at which hCG was detectable and the level of hCG detected on that day were determined with the hCG CIA. The first day at which hCG was detectable was defined as the first day that the daily hCG concentration, as measured by hCG CIA, exceeded a sliding 5-day average baseline plus 2 SD for two of three consecutive days beginning with FSH Days +1 to +5. Group average, SEM, and CV were calculated for the cycle and FSH days of hCG detection, the corresponding hCG concentrations, the 5-day average hCG baseline concentrations, and the incremental change between the hCG concentrations on the day of detection and the 5-day average baseline concentration.
hCG B:I Profile Analysis
Following analyses with the hCG CIA, up to 12 daily samples were selected from each conceptive cycle's available sample set for hCG B:I profile analysis with the intact hCG IEMA and the LH/CG cellular bioassay beginning on the first sample day with measurable hCG as determined by the hCG CIA. The hCG IEMA and LH/CG cellular bioassay were used to measure IhCG and BhCG levels, respectively, so that the results obtained for the present study would be comparable to those originally published by Ho et al. [18] for circulating hCG isoforms. Daily BhCG and IhCG levels were aligned by FSH day and used to calculate daily hCG B:I ratios for each cycle by dividing the former value by the latter. FSH Days +10 to +19 were selected for the analysis, because they spanned the same period of early pregnancy already used successfully to analyze the hCG B:I ratio in the serum of normal pregnancies [18]. Group average, SEM, and CV were calculated for daily IhCG, BhCG, and hCG B:I ratio values, and plots of the group average values were prepared for each measurement.
To test our hypothesis that the rate at which the hCG B:I ratio increases is characteristic in surviving and failing pregnancy outcomes, two 5-day analysis windows (FSH Days +10 to +14 and FSH Days +15 and +19) were used. Five-day analysis windows were used for the previous analysis of daily serum samples [18]. In the present study, the use of two consecutive 5-day analysis windows permitted a preliminary assessment regarding the effect of analysis window timing on the sensitivity of the hCG B:I ratio biomarker. The rate at which the hCG B:I ratio increased in the conceptive cycles in the present study was evaluated with linear-regression analysis. Linear slopes for daily IhCG, BhCG, and hCG B:I ratio levels were calculated for all individual cycles that resulted in LB, CAB, and EPL that had at least 4 of 5 days within a given 5-day analysis window. Each cycle's individual linear slope was then used as a metric to compare the rates of change for daily IhCG, BhCG, and hCG B:I ratio levels between outcome groups within that 5-day analysis window. Transverse mean (TM) hCG B:I ratio levels, defined as an average level across a 5-day analysis window, also were calculated for the each cycle. The TM was then used as a metric to determine whether the rate at which the hCG B:I ratio levels increase is more characteristic of pregnancy outcome compared with the average daily level.
Laboratory Analyses: Automated Hormone Assays
Hormone assays for FSH and hCG previously formatted for use on microplates have been adapted for the automated ACS-180 Autoanalyzer (Bayer Diagnostics). Because of the sample volumes required for the present study, all hormone measurements of daily urine samples were performed in singleton, and each assay included standards, samples, and internal controls.
The peptide hormone subunit FSHB was measured with a two-site FSH CIA with the following previously described reagent pair: FS2 mouse-anti-ßFSH-dimethyl-acridinium ester (DMAE) primary monoclonal antibody and R9715 rabbit-anti-FSHB secondary polyclonal antiserum [20]. Briefly, in this immunoassay configuration, disassociated FSHB subunit in the patient sample (boiled undiluted urine), standard, or internal control is bound by DMAE-labeled primary anti-FSHB monoclonal antibody and polyclonal rabbit anti-FSHB secondary antibody. The primary anti-FSHB monoclonal antibody:FSHB:polyclonal rabbit anti-FSHB antibody complex is then bound to monoclonal mouse anti-rabbit antibody, which is covalently coupled to paramagnetic particles (Qiagen). Reagents and hormones not involved in the immune complex are aspirated, whereas the immune complex is retained by the application of a magnetic field. Retained DMAE in the immune complex is then induced to emit light by the addition of hydrogen peroxide and sodium hydroxide. The amount of urinary hormone in the patient sample is related directly to the amount of relative light units detected by the system. The interassay CVs for low, medium, and high internal control samples were 41.55%, 16.41%, and 12.75%, respectively.
The peptide hormone hCG was measured using a two-site hCG CIA with the previously described reagent pair, B109 mouse-anti-intact hCG primary monoclonal antibody-DMAE and R76 rabbit-anti-hCG secondary polyclonal antiserum [17, 21], following the same assay configuration and scheme as described for the FSH CIA in the previous paragraph. The interassay CVs for low, medium, and high internal control samples were 19.01%, 5.76%, and 7.04%, respectively.
Each selected subject's daily urine samples were thawed, after which sufficient volumes (<1.0 ml) were aliquotted, returned to –20°C, and transported to our laboratories at the University of California, Davis, for subsequent hormone measurements. All daily urine samples from each cycle were first analyzed with an automated FSH CIA and companion colorimetric CR assay to define the beginning of each cycle's implantation window. Counting backward from the last day of collection, up to 20 samples were then selected for analysis with the hCG CIA. Daily FSH data from days with CR levels of less than 0.2 mg/ml were censored, and the remainder were indexed with daily CR concentrations to adjust for urine concentration. Daily hCG CIA data were not indexed with CR, both because these data were employed to detect the earliest appearance of hCG in a conceptive cycle and because elevations resulting solely from day-to-day fluctuations in urine concentration could be confused with the initial rise of hCG. Unindexed daily hCG concentrations with low CR were not censored, but hCG values less than the approximate theoretical sensitivity of the hCG IEMA (0.05 ng/ml) were replaced with 0.05 ng/ml.
hCG B:I Profile Hormone Assays
Selected daily urine samples were diluted 1:1, 1:8, 1:32, and 1:128 in Dulbecco minimal essential medium (DMEM) with 1% fetal bovine serum (Gibco), and all measurements of the prediluted daily urine samples were performed in singleton. Each sample's hCG measurements were selected from the first dilution in the series with an absolute concentration that did not exceed the linear portion of each immunoassay or cellular bioassay standard curve.
The LH/CG cellular bioassay used was based on the cloned human fetal kidney cell line 293 that had been permanently transfected with hLH/hCG receptor and luciferase reporter gene cDNA [16]. Biologically active hCG in a sample incubated with the cells in culture is detected, following hCG binding to the LH/hCG receptor, by an increase in cAMP-mediated protein kinase A activity, the activation and expression of luciferase genes, and the subsequent luciferase substrate conversion, with a luminometer. Each cellular bioassay included standards and two internal control levels, with interassay CVs of 17.81% and 27.2% and a previously reported mean sensitivity of 0.05 ng/ml [17].
The hCG IEMA used was based on a monoclonal primary antibody that specifically recognizes the intact hCG heterodimer (B109) and a biotin-conjugated secondary monoclonal antibody (B108-biotin) that specifically recognizes the hCG ß subunit [21]. Alkaline phosphatase streptavidin (APS; Zymed) is incubated with the biotin in the B109:hCG:B108-biotin complex so that the amount of intact hCG heterodimer in a microplate well is directly proportional to the amount of B109:hCG:B108-biotin:APS complex, as determined by APS substrate conversion measured with an optical microplate reader. Each IEMA included standards and two internal control levels, with interassay CVs of 9.02% and 11.43% and a previously reported mean sensitivity of 0.035 ng/ml [17].
Daily immunoreactive hCG (IhCG) and bioactive (BhCG) concentrations were indexed with daily CR concentrations to adjust for urine concentration, and data from days with CR levels of less than 0.2 mg/ml were censored before normalization. Daily hCG B:I ratios were calculated by dividing daily BhCG measurements by their corresponding IhCG measurements. Because this was a ratio of two measurements, daily hCG B:I ratios with corresponding low CR levels were not censored. Daily IhCG and BhCG concentrations less than the approximate theoretical sensitivity of the hCG IEMA (0.05 ng/ml) were replaced with this value.
Data Analysis
All column statistics, ANOVA, Student t-tests, and figures were prepared with Prism v.4 for Windows (GraphPad Software).
RESULTS
Samples from this data set have been analyzed in previously published studies of toxicant-induced EPL [8, 9]. The data reported here were obtained from the conceptive menstrual cycles resulting in LB of 30 of 32 randomly selected subjects enrolled in a research protocol to study EPL. The hormone data from two subjects were excluded from final analysis. In the first case, the subject's diary data were not consistent with hormonal landmarks, and in the second case, no samples were available for the analysis window between FSH Days +10 and +19. The timing of the day of ovulation was identified with the CR-indexed FSH peak in 29 of 30 conceptive menstrual cycles resulting in LB (Table 1). The conceptive cycle for subject 17 had more than 3 of the 5 days surrounding the CR-indexed FSH peak value censored, because those days had CR levels of less than 0.2 mg/ml. In this case, the indexed FSH peak could not be identified with certainty, so the unindexed FSH peak was used in its place for the alignment of subsequent data. Five of five conceptive cycles that resulted in CAB and five of six conceptive cycles that resulted in EPL were included in the final data analysis in this report. One conceptive cycle resulting in an EPL was excluded from these analyses, because it had no daily hCG CIA levels that exceeded 0.1 ng/ml. Therefore, it was concluded that there would be insufficient analyte for the IhCG and BhCG assays after dilution with DMEM.
|
For conceptive cycles resulting in LB, CAB, and EPL, the CR-indexed FSH peak was observed to occur between Cycle Days 9 and 19, 8 and 17, and 11 and 38, whereas the average peak CR-indexed FSH concentrations ranged from 0.59 to 4.15, 0.52 to 3.48, and 0.37 to 1.81 ng FSH/mg CR, respectively (Table 1). The estimated timing of implantation was identified in all conceptive menstrual cycles with the hCG CIA and was reported as either a cycle day or, relative to the day of ovulation, as an FSH day (Table 1). The cycle days of hCG detection were observed to range between Cycle Days 18 and 33, 16 and 27, and 21 and 50, whereas the corresponding FSH peak-aligned days of hCG detection were observed to range between FSH Days +6 and +14, +7 and +11, and +10 and +14 for conceptive cycles resulting in LB, CAB, and EPL, respectively (Table 1). For each outcome group, the difference between the maximum and minimum values for the day of hCG detection was reduced when the data were aligned by their corresponding FSH days. For this reason, estimates for the timing of implantation were relative to the CR-indexed FSH peak. The average baseline hCG value for the hCG CIA was 0.052 ng/ml for the conceptive cycles in all three outcome groups, whereas the average incremental rise in the concentration of hCG observed on the day of detection was observed to be 0.23, 0.13, and 0.21 ng hCG/ml for conceptive cycles resulting in LB, CAB, and EPL, respectively (Table 1).
When the FSH and hCG CIA data were plotted and compared, there was considerable overlap between outcome groups with regard to the timing of ovulation (Fig. 1A) and the magnitude of the FSH surge (Fig. 1B), and the incremental rise of the hCG level over baseline on the day of first detection (Fig. 1C). The latter observation was consistent with our previously reported data [22]. Although the timing of implantation was similar for pregnancies that resulted in LB and CAB, a significant difference (Student t-test, P < 0.05) was observed when the data for pregnancies that resulted in LB and EPL were compared (mean ± SEM: 9.933 ± 0.43 days vs. 12.20 ± 0.66 days, respectively) (Fig. 1D).
|
Daily IhCG, BhCG, and hCG B:I ratio data for 30 conceptive menstrual cycles resulting in LB, 5 conceptive menstrual cycles resulting in CAB, and 5 conceptive menstrual cycles resulting in EPL were aligned by the estimated day of ovulation (FSH Day 0). Each outcome group's average levels of each analyte were then plotted for FSH Days +10 to +19. The average daily pattern of hCG excretion for cycles resulting in LB was characterized by a consistent rise of IhCG levels for FSH Days +10 to +19 (Fig. 2A). For this group of pregnancies, the daily pattern of BhCG excretion was observed to be similar to that of IhCG. No significant differences were found when daily IhCG and BhCG measurements were compared (Student t-test, P > 0.05). For the CAB outcome group, the average daily pattern of IhCG excretion was characterized by IhCG levels that rose from FSH Days +10 to +17 and declined between FSH Days +17 and +19 (Fig. 2B). The daily pattern of BhCG excretion was observed to be similar to that of IhCG for the CAB group. The average daily pattern of IhCG excretion for cycles resulting in EPL was characterized by inconsistently rising levels between FSH Days +10 and +19. For the EPL group, the daily pattern of BhCG excretion was observed to be similar to that of IhCG except between FSH Days +10 and +13, when it appeared to be lower (Fig. 2C). Although the inconsistencies in the IhCG and BhCG profiles were especially evident for the EPL outcome group, the only significant differences in either of these profiles for pregnancies that resulted in either CAB or EPL were observed for FSH Day +14 (Student t-test, P < 0.05). This result is not consistent with trends that suggest daily BhCG levels generally were lower than their IhCG counterparts and, therefore, should be interpreted conservatively, because limited data (two or fewer data points per analyte) prevented comparisons on FSH Days +18 to +19 and +10 to +12 for the CAB and EPL outcome groups, respectively. As well, the only significant differences observed between pregnancies that resulted in LB and pregnancies that resulted in CAB or EPL in the daily IhCG and BhCG measurements were observed for FSH Day +14, when the average daily levels of both measurements were lower for the failing pregnancies (Student t-test, P < 0.05). As before, this result is not consistent with the trends that suggest IhCG and BhCG levels generally were higher for pregnancies that resulted in LB than for those that resulted in EPL or CAB and also should be interpreted conservatively, because the limited data for the CAB and EPL outcome groups prevented comparisons on the previously mentioned analysis days.
|
The longest continuous rise in the pattern of the daily average hCG B:I ratio for the LB outcome group was observed to occur between FSH Days +15 and +18 (Fig. 2A), whereas during the same time period, inconsistent patterns were observed for the CAB and EPL outcome groups. Based on this observation and others from our previous study [18], we concluded that the analysis window bounded by FSH Days +15 and +19 could represent an area with measurable differences in excreted hCG analytes between surviving and failing outcome groups, and we used it as a primary focus for the further characterization of individual conceptive cycles. On the other hand, differences were observable between surviving and failing pregnancies during the time period bounded by FSH Days +10 and +14 that suggested this window warranted further analysis for the purpose of comparing the reliability of the hCG B:I ratio biomarker for pregnancy outcome at two distinct points in the greater periimplantation period.
Seventeen conceptive cycles that resulted in LB, four cycles that resulted in CAB, and one cycle that resulted in EPL that met criteria described in Materials and Methods were carried forward for individual linear-regression and TM analyses of the IhCG, BhCG, and hCG B:I ratio levels between FSH Days +10 and +14 (Table 2). Additionally, 15 conceptive cycles that resulted in LB, 2 cycles that resulted in CAB, and 4 cycles that resulted in EPL that met criteria described in Materials and Methods were carried forward for individual linear-regression and TM analyses of the IhCG, BhCG, and hCG B:I ratio levels between FSH Days +15 and +19 (Table 2).
|
Figure 3, A–C, provides a graphic, side-by-side comparison of the linear-regression slopes from individual conceptive cycles for the analysis window at FSH Days +10 to +14. This figure shows complete overlap of the linear slope values for IhCG, BhCG, and hCG B:I ratio levels of conceptive cycles resulting in CAB and EPL by those linear slope values for conceptive cycles resulting in LB. No statistically significant differences between outcome groups by ANOVA (P > 0.05) were observed for any parameter. When the analysis window at FSH Days +15 to +19 was used, the linear slope values for IhCG and BhCG of conceptive cycles resulting in CAB and EPL were completely overlapped by the linear slope values for conceptive cycles resulting in LB (ANOVA, P > 0.05). However, 79% of the hCG B:I ratio slope values for the LB group were higher than the highest slope for the CAB outcome group; this difference was not statistically significant (Student t-test, P > 0.05) (Fig. 3, E–G). For the same analysis window, a single hCG B:I ratio slope value for the EPL group (subject 37; daily hCG B:I ratio, 0.35) was observed that was higher than all the slopes for the LB and CAB outcome groups (Table 2 and Fig. 3G). When this slope was censored and the data reanalyzed, 71% of the hCG B:I ratio slope values for the LB group were higher than the highest remaining slope for the EPL outcome group, and this difference was statistically significant (Student t-test, P < 0.05). The TM hCG B:I ratios between all the outcome groups were then compared, and we observed complete overlap when the analysis window at either FSH Days +10 to +14 or FSH Days +15 to +19 was used (Fig. 3, D and H).
|
Subjects 22 and 29 contributed outlying IhCG and BhCG slopes within the LB outcome group (Table 2). In each case, a single outlying IhCG or BhCG data point greater than 400% of the daily mean value was responsible. To assess whether the outlying data contributed by both of these subjects unduly biased the group IhCG and BhCG profiles for the LB outcome group, these values were censored and the daily group mean values recalculated. After the recalculation, the average daily IhCG and BhCG concentrations were reduced for FSH Days +18 and +19 and for FSH Day 18, respectively, and the average daily pattern of hCG excretion for cycles resulting in LB was characterized by a consistent rise in IhCG levels for FSH Days +10 to +19. The daily pattern of BhCG excretion was observed to be similar to that of IhCG; no significant differences were found when daily IhCG and BhCG measurements were compared (Student t-test, P > 0.05).
DISCUSSION
For the purposes of comparison, it was necessary to use the same immunoassay and cellular bioassay originally employed by Ho et al. [18] for the hCG B:I ratio analysis, but the expense and complexity of these analyses necessitated that only those samples within a confirmed conceptive cycle's periimplantation period be carried forward for analysis. We therefore used a physiologic implantation window bounded by estimates for the timing of ovulation and implantation for each cycle in the study to determine which samples within each cycle to analyze with the hCG IEMA and the LH/CG cellular bioassay and, later, for final data alignment. Automated CIAs for FSH and hCG were used for the purpose of obtaining these estimates in this report.
When we compared the estimated timing of implantation between pregnancy outcome groups, pregnancies that resulted in EPL were found to implant significantly later than those that resulted in LB, but complete overlap was observed in the data. Evidence already indicates that the risk of pregnancy loss increases with later implantation times relative to the timing of ovulation [23, 24], and although our findings must be interpreted with caution because of the limited number of cycles analyzed, they do support these previous observations.
The hCG metabolites in urine are known to include intact hCG heterodimer, enzymatically nicked heterodimer, free CGA and CGB subunits, and various enzymatic fragments, all with varying degrees of carbohydrate microheterogeneity [15]. All the aforementioned modifications of hCG's polypeptide backbone have been reported to attenuate binding to the human LH/CG receptor and its downstream signal [17]. Additionally, the urine matrix varies greatly from individual to individual and from day to day in both concentration and composition, because it is an individual product of hourly intake and outflow. Previous attempts to measure urinary hCG metabolites with a less robust LH/CG bioassay cell line were limited by a urine matrix effect that interfered with direct measurement (data not shown), but the cell line used for the present study permitted the measurement of daily urine samples that were prepared by simple dilution in tissue culture medium.
All the conceptive cycles from surviving and failing pregnancies that were analyzed for this report had measurable levels of BhCG isoforms in daily urine samples collected during the analysis window originally used by Ho et al. [18] to survey the original characterizations of hCG isoforms in daily blood samples. The observed differences in the pattern of the secretion of BhCG isoforms between surviving and failing pregnancies were greatest during the periimplantation period that overlaps the window bounded by FSH Days +10 and +19, so further development of an early biomarker for pregnancy outcome applicable to field studies based on these observations may depend on the ability of investigators to monitor this critical period of time with daily urine samples.
The original observations by Ho et al. [18] led to the hypothesis that in normal pregnancies, an early hCG isoform with reduced bioactivity was produced at the time of implantation, and that its synthesis was gradually eclipsed by that of a later hCG isoform with greater relative bioactivity within the first 2 wk postimplantation, whereas in abnormal or failing pregnancies, the rate and/or magnitude of synthesis of the putative later hCG isoform was deficient. In the present study, we observed consistent and robust increases in average daily IhCG and BhCG levels across the analysis window for the conceptive cycles that resulted in LB. More importantly, we observed inconsistent increases with pronounced periods of decline in the average daily IhCG and BhCG levels across the analysis windows for CAB and EPL. However, the more steeply rising BhCG levels as compared to IhCG levels observed in daily periimplnatation blood samples by Ho et al. [18] were not observed for any of the outcome groups with the measurements of urinary metabolites during the same time period.
Many of the structural features necessary for the bioactivity of hCG molecules have been described in detail [25–28], but to our knowledge, the differences in the bioactivities of circulating and urinary hCG molecules during the periimplantation period of early pregnancy have not been compared directly. It is not presently known what role the differences in metabolism, clearance, and excretion for the different hCG isoforms may have played, but each likely had an effect on the urinary IhCG and BhCG profiles that we observed. Further metabolism and structural studies are certainly needed.
The greatest differences between pregnancies resulting in LB and those resulting in CAB and EPL were observed when slopes for the hCG B:I ratio were used as the analysis parameter for the 5-day window bounded by FSH Days +15 to +19. This is consistent with the original observation by Ho et al. [18] that the rate at which the hCG B:I ratio increased was steeper for surviving pregnancies than for those that ultimately failed. There appeared to be differences in the hCG B:I ratio profiles between surviving and failing outcome groups, but we previously reported that periimplantation hCG levels of early pregnancy are highly variable in blood and urine and do not correlate with pregnancy outcome in our clinical study population when only immunoassay is used [22]. Taken together, these data suggest that although the patterns of hCG secretion observed in the daily periimplantation blood samples of surviving and failing pregnancies by Ho et al. [18] were different from the analogous patterns observed for urinary hCG metabolites, critical information relating to pregnancy outcome may be present in blood and urine samples collected during the same time period when bioactivities and immunoreactivities are measured in concert.
The LH/CG cellular bioassay is presently the only method that can provide specific information about the in vivo bioactivity of hCG molecules in biological samples. However, this method is technically demanding, costly, and at present, not transportable to the field. Therefore, the development of an immunoassay specific to the structural features of hCG isoforms with full or attenuated bioactivities may be necessary for the widespread application of a biomarker based on hCG bioactivity. However, the observed differences in the daily periimplantation patterns of hCG secretion and excretion are still poorly understood, and the responsible hCG isoforms and their exact structural features have not yet been identified.
Much attention has been paid to the relationship of carbohydrate heterogeneity to the bioactivity of hCG isoforms [29], but evidence is mounting that the physical conformation of the CGA component of the hCG heterodimer is a key factor in LH/CG receptor signal transduction [25–28]. This has led us to hypothesize that the physical conformation of the CGA subunit may represent a novel avenue to approach the development of specific immunochemical reagents to measure a putative early isoform of hCG with low bioactivity. In 2002, Hsu et al. [30] and Nakabayshi et al. [31] identified a heretofore-unknown gene for a putative glycoprotein subunit (CGA2) that was subsequently found to associate strongly with CGB in a heterodimer with reduced bioactivity. This observation was consistent with the hypothesis of Ho et al. [18], because it described a putative isoform of hCG with reduced or null bioactivity as compared to mature hCG that may be produced early and is eclipsed by a later isoform of hCG with full bioactivity. If this were true, then immunoassay reagents could be prepared to measure specifically the hCG isoform with reduced bioactivity, thus making the LH/CG cellular bioassay unnecessary for field application of an early biomarker for pregnancy outcome based on the function of hCG as a tropin. Ultimately, this potential immunoassay could be adapted to an automated platform and used in conjunction with the previously automated hCG CIA.
The present results suggest that the hCG B:I ratio may be useful as an early biomarker, but several limitations of this study must be considered in the final analysis. First and foremost were the limitations imposed by the size of the study population. The cycles used for the present study originally were collected for a large prospective study between 1996 and 1998, without the particular needs of the present study in mind [8]. Since collection, many of the daily urine samples in the study sample set have been depleted, so whereas our cycle selection was random, the availability of a volume of urine that was sufficient to complete our analyses was the final selection criterion. Also, in field studies of the type for which the samples originally were collected, CAB and EPL are rare events, the paucity of which invariably affected the cycle selection for the present study. Despite these limitations, the present data provide evidence that the analysis of multiple urinary metabolites of hCG may provide additional information related to pregnancy outcome and support further studies in this area.
Daily hormone data from the hCG CIA assay, the hCG IEMA, and the LH/CG cellular bioassay were replaced with the approximate theoretical sensitivity for all the assays (0.05 ng/ml) when the daily data were less than this value before indexing with CR. Although it is standard practice to use theoretical or practically derived sensitivities or limits of detection, it certainly requires justification anytime it is used in a situation when the measurement of low levels of a given analyte is necessary. In the case of the hCG CIA, 30 cycles that resulted in LB had an average baseline level of 0.052 ng/ml, whereas the average hCG level on the day of detection was 0.28 ng/ml, which is more than 500% of the baseline value. This is less of a concern for the hCG IEMA and the LH/CG cellular bioassay, because low values were rare during the time period of FSH Days +10 to +19 and had little effect on group averages, slopes, and TMs. All the aforementioned metrics are reliable, because all are products of multiple data points. The interassay CVs for the low internal controls for the FSH and hCG CIA assays both exceeded 15%. Because the FSH assay was used to determine the timing of the FSH peak, the quantification of hormone levels that were comparable to the low internal control was less important than the quantification of peak levels that were comparable to the middle and high internal controls, which each had a lower CV for the analyses described in this report. In contrast, the CV for the baseline hCG levels in the daily urine samples analyzed was observed to be 9.31%, and in each cycle analyzed for the present study, the hCG level on the day of detection was greater than the average baseline levels plus 2 SD.
To the best of our knowledge, this is the first study to characterize of the bioactivity of urinary metabolites of hCG in the daily periimplantation urine samples of surviving and failing pregnancies. Taken together, the present data provide evidence that suggests urinary metabolites of hCG may provide additional information relating to pregnancy outcome when analyzed for multiple isoforms and that a biomarker based on this information might be developed for population-based studies and applied with a physiologic implantation window. Although limited in size and statistical power, the current data support further studies in this area.
FOOTNOTES
1 Supported by the Superfund Basic Research Program (P42ES04699) and NIEHS (P01ES06198). 
2 Correspondence: Bill L. Lasley, Center for Health and the Environment, University of California, Davis, CA 95616-8615. FAX: 530 752 5300; bllasley{at}ucdavis.edu
Received: 10 October 2005.
First decision: 6 November 2005.
Accepted: 7 March 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  P.A. Nepomnaschy, C.R. Weinberg, A.J. Wilcox, and D.D. Baird Urinary hCG patterns during the week following implantation Hum. Reprod., February 1, 2008; 23(2): 271 - 277. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
