| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Southwest National Primate Research Center (S.D.T., D.G.L.), San Antonio, Texas 78245; Department of Psychology and National Primate Research Center (T.E.Z.), University of Wisconsin, Madison, Wisconsin 53715; and American College of Obstetricians and Gynecologists (M.P.), Washington, D.C. 20024
Address all correspondence and requests for reprints to: Dr. Suzette Tardif, Southwest National Primate Research Center, P.O. Box 760549, San Antonio, Texas 78245-0549. E-mail: stardif{at}icarus.sfbr.org.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
With a relatively high fertility and short life span, marmoset monkeys (Callithrix jacchus) offer an opportunity to develop useful primate models of prenatal nutritional effects on birth condition and adult disease risk. Establishment of such models will be dependent upon the development of reliable means to induce prenatal nutritional change and reliable assessment of placental/fetal function. We have examined the effects of maternal energy restriction during pregnancy in the common marmoset (2) and found that one particular restriction regimea modest energy restriction in midpregnancyreliably induces pregnancy loss 1147 d (mean = 25) after initiation of restriction.
The pregnancy losses associated with this midterm energy restriction could be the result of a number of different mechanisms, and some of those mechanisms might be distinguished by comparison of maternal endocrine changes during restriction. For example, energy restriction could act as a stressor resulting in increasing maternal cortisol concentration and increasing placental CRH production (or decreasing production of CRH binding protein). Increased CRH would, in turn, result in increased fetal adrenal production of dehydroepiandrosterone (DHEA) sulfate. This would lead to increased estradiol production by the placenta, triggering changes in prostaglandin function, leading to onset of contractions and delivery. Conversely, energy restriction could result in a decrease in placental CRH production, which might adversely affect fetal adrenal development (3) or, in hypoxic conditions, favoring reduced IGF-I and increased IGF binding protein, ultimately leading to reduced placental growth and fetal demise. Either of these cases would be associated with decreasing maternal serum concentration of cortisol and estradiol.
Because of their small body size and, therefore, limited availability of sequential blood samples, determination of changes in urinary concentrations of hormones during normal and restricted pregnancies is of particular value in this species. Excretion of maternal estrogens has been used to assess placental/fetal viability in women (4, 5). Low levels of urinary estrogens have been used to indicate placental insufficiency associated with retarded intrauterine growth (4). Circulating and urinary levels of maternal cortisol show a typical pattern in primates of basal levels during the first half of pregnancy with an abrupt sustained elevation during the last half of pregnancy (see Ref. 6 for review). The sustained elevated cortisol levels are attributed to the development of the transitional zone of the fetal adrenal glands (7) and allow for a method to monitor fetal function through their adrenal production of cortisol.
This report describes changes in maternal excretion of cortisol, estrogens, and chorionic gonadotropin (CG) in normal, term pregnancies vs. energy-restricted, aborted pregnancies. Our goals were 1) to determine which form of estrogens (estradiol or estrone as free steroids and as those removed from conjugates) would best reflect viability of the placenta as a steroid synthesizing organ; 2) to determine whether the energy restriction regime appeared to be acting as a classical stressor as evidenced by increasing cortisol and, subsequently, estradiol concentration or, alternately, if cortisol and estradiol levels were reduced in restricted pregnancies, associated with mechanisms resulting in restriction-induced impairments in placental function; 3) to determine whether CG levels would indicate a reduction before pregnancy loss indicative of disruption of placental growth. CG was examined as an independent marker of placental differentiation and function.
| Materials and Methods |
|---|
|
|
|---|
Six female common marmosets were followed through two pregnancies: a normal full-term pregnancy and a pregnancy in which they were energy restricted. Females were multiparous with previous successful pregnancy outcomes. As a moderate restriction, subjects were restricted to a known caloric intake, based on their prepregnancy weight, which was 75% of the expected intake for a female of their body mass. Subjects were fed either of two purified diets, specifically formulated for these studies (8). Restricted females were fed a 25% protein diet, resulting in their protein consumption being roughly equivalent to that of ad-libitum fed females who were fed the 15% protein diet. The diets were formulated such that all other nutrients in the diet (e.g. calcium, iron, folic acid, and vitamins B6 and B12) were in sufficient excess that a 75% restriction would still likely provide adequate nutrition. Control of food intake was accomplished by separating females on one side of the double-unit home cage. Females had visual, olfactory, auditory, and limited tactile access to the other members of their family group through a mesh door dividing the two units. Diet restriction began at estimated gestational ages ranging from 60 to 72 d (mean of 66.9). All females were given an ultrasound exam every 2 wk to determine the stage of pregnancy and the size and condition of embryos/fetuses. Gestational age was estimated by embryonic crown-rump lengths taken during the early, rapid phase of growth in crown-rump length, between d 55 and 70. This measure can be used to consistently estimate delivery dates (9). Additional details on the energy restriction regimen are found in (2).
Morning-void urine was collected from the females several times per week throughout the 143-d gestation (10). Urine was collected by separating the female on one side of the double-unit home cage after placing plastic under the cage (cages were elevated 18 inches above the floor). After the female urinated, urine samples were pipetted into small cryovials and stored frozen at 20 C. Samples were shipped to the National Primate Research Center (University of WisconsinMadison) to the Assay Services laboratory.
Hormone assays
Samples were measured for estrogens, cortisol, CG, and creatinine in matched samples between the two pregnancies. All hormones were indexed by the measurement of creatinine to adjust for fluid variations between samples.
Urine samples were analyzed directly for cortisol by enzyme immunoassay using the technique described in Ref. 11 . Marmoset urine was diluted 1:1000 with assay buffer, and 50 µl was added to duplicate wells. The antibody (R4866; developed by G. Stabenfeldt, University of California, Davis, CA) cross-reacts 60% with cortisone, a major metabolite of cortisol in marmosets, 2.5% with corticosterone, and less than 1% with other steroids. This method has been routinely used for both marmosets and tamarins to reflect cortisol excretion in urine and reflects reproductive condition (12). Serial dilutions of pooled marmoset urine were parallel to the cortisol standards: t = 1.1, P > 0.05, n = 9. Accuracy for added marmoset urine to the standard curve points was a mean of 94.42 + 3.41 SD, n = 6. Intraassay coefficients of variation for a low and high pool were 2.4 and 1.0%, respectively, and interassay coefficients of variation for the same pools were 14.6 and 14.3%.
For estrogen measurement, the urine samples from pregnant marmosets were selected from the gestational days of 4060 for early pregnancy, 75110 for midpregnancy, and 120140 for late pregnancy. For each gestational phase (early, mid, and late), five samples were pooled by taking 100 µl of urine from each sample and combining for sample preparation. Samples were separated for free estrogens and conjugated estrogens by extracting the urine with 5 ml ethyl ether after the addition of 500 µl H2O. The unconjugated steroids were pipetted off in the solvent phase, dried under air, and resuspended in 1 ml 30% methanol. The aqueous phase was sent through solvolysis to remove the conjugates from the steroids by the technique reported in Ref. 13 . Basically, for each 1-ml aqueous sample, 100 µl saturated NaCl, 50 µl 2.5 M H2SO, and 4 ml ethyl acetate were added to the sample, vortexed, and incubated for 2 h at 60 C. After incubation, the samples are vortexed for 5 min and centrifuged for 2 min at 1000 x g, and then the solvent phase was pipetted off, dried, and resuspended in 30% methanol.
Both the free and conjugated fractions were then sent through solid phase extraction to purify the sample before HPLC. The solid phase extraction (60 mg/3 ml; Strata X, Phenomenex, Torrance, CA) was performed as indicated on the package directions with the following exceptions: the 1-ml samples were added as an aqueous solution with 30% methanol. Samples were washed with 1 ml 20% methanol before elution with 1 ml methanol. Purified samples were then dried, transferred to HPLC vials, and resuspended in 20 µl of acetonitrile:water (50:50).
HPLC was performed to separate estradiol from estrone before RIA by the technique reported in Refs. 6 and 14 with the following changes: samples were injected in 20 µl and run isocratically for 30 min at 40/60% acetonitrile/water. From each sample, fractions were collected for estradiol eluting at 11.9 and estrone eluting at 19.1 min. Retention times (in minutes) varied little; coefficients of variation for estradiol was 0.01 and for estrone was 0.01. Collected fractions were dried and reconstituted in 500 µl ethanol. RIAs were performed for each separated estrogen by the methods reported in Ref. 15 . Estradiol samples were diluted 1:10 (free fractions) and 1:100 (conjugated fraction). Estrone samples were diluted 1:100 for both free and conjugated fractions.
CG was measured according to the methods reported in Ref. 16 for measuring gonadotropins in Callitrichid monkeys and validated for the common marmoset. This assay reflects both LH secretion and CG in Callitrichids but uses human CG as the radioligand and standards (hCG; CR-127). Intra- and interassay coefficients of variation for this assay are 8.17 and 14.14, respectively.
Creatinine measurement was taken using a spectrophotometric assay using the Jaffe reaction method for measuring creatinine as described in Ref. 11 .
Comparisons were made between normal, term pregnancies and energy-restricted, aborted pregnancies by paired t tests. Potential effect of litter size (determined by either ultrasound detection of heartbeat number or by number of term or aborted infants) was examined to ensure that any differences between term and aborted pregnancies were not due to different litter sizes.
| Results |
|---|
|
|
|---|
Figure 1
provides profiles of cortisol over the gestational weeks (1 - 21 wk). Full-term pregnancies showed the typical primate profile of basal levels of cortisol for the first half of pregnancy with sustained elevations of cortisol throughout the latter half of pregnancy beginning around wk 10. Aborted pregnancies had lower cortisol levels and a delayed increase in cortisol. Comparison of average cortisol levels during the last 2 wk of the restricted pregnancy and matching weeks from a full-term pregnancy for each female showed significant differences in cortisol levels. Mean cortisol levels for the full-term pregnancies (77.36 ± 16.6, n = 6) were significantly higher than matched cortisol levels during the restricted pregnancy (60.91 ± 12.84, n = 6), by matched t test (t = 2.7, P = 0.04). Cortisol levels were on average 79.71 ± 5.89% of normal levels (range = 55.29, 97.64) in the energy-restricted pregnancy.
|
|
|
Secondly, because energy restriction might cause an increase in excreted creatinine, due to increased metabolism of protein as an energy source, we compared the creatinine concentrations in the two groups. For those samples used to compare cortisol concentrations, there was no difference between term and aborted pregnancies (0.643 µg/ml vs. 0.772 µg/ml, paired t test = 0.896, P = 0.41). For those samples used to compare average estrogen and CG concentration, there was a significant difference in the creatinine levels (0.548 vs. 0.844, paired t test, T=2.99, P = 0.03). We therefore compared concentrations of free estradiol and CG per milliliter of urine, to ensure that the differences observed were not simply due to the difference in creatinine. Raw concentrations of CG and free estradiol displayed the same difference as that seen in concentrations gauged per microgram of creatinine; i.e. concentrations were significantly lower in aborted pregnancies than in term pregnancies (CG: 51.57 vs. 15.57; paired t test = 4.963, P = 0.016; free estradiol: 293.655 vs. 109.14; paired t test = 2.15, P = 0.04).
| Discussion |
|---|
|
|
|---|
Additional support for the hypothesis that energy restriction led to impaired placental function is found in the reduced urinary CG concentration in pregnancies that would ultimately abort. CG is produced by the embryonic trophoblastic cells beginning during the preimplantation interval and is the signal that rescues the corpus luteum at the end of the ovarian cycle when pregnancy occurs (1). The pattern of secretion of CG in marmosets is similar to that of humans, with a marked rise and fall occurring during early-mid pregnancy (17). However, CG elevation to peak production in marmosets is a more gradual process, most likely associated with the prolonged, largely quiescent period with slow cell division and differentiation from implantation, around d 11, until around d 50, at which point developmental processes speed up, with organogenesis completed by d 80 (18). In humans, CG is an intrinsic regulator of trophoblast differentiation and function (19). Concentration of CG is related to the number of cytotrophoblast cells (20) and, as such, may serve as a marker for placental differentiation and function in early to mid pregnancy. The reduced excretion of CG in the energy restricted mothers, therefore, further supports the contention that placental differentiation or function was impaired by some aspect of the restriction.
Maternal urine provides hormonal markers of both placental and fetal function before abortion or reabsorption of the placental and fetal tissues in the common marmoset. The concentrations of steroids in the maternal circulation or urine are the result of a complex interaction between the maternal neuroendocrine system, the placenta, and the fetus (21). For example, production of estradiol by the placenta is dependent upon availability of C-19 precursor molecules produced by the fetal adrenal, while, at the same time, maternal estradiol concentrations affects interconversion of cortisol and cortisone, therefore affecting the exposure of the fetal adrenal to cortisol, regulating the fetal hypothalamo-pituitary-adrenal axis and the onset of cortisol production by the fetal adrenal (22). These complex interactions of the placenta and fetus are such that our limited sampling period for the estrogens, cortisol, and CG did not allow us to determine whether the placental and fetal function was differentially affected by energy restriction. However, they do provide us with information that indicates that energy restriction leads to a fairly rapid decrease in all of these markers indicating loss of placental and fetal function.
Two issues are of interest in examining marmoset urinary estrogens. The first is that the reported predominant estrogen is dependant upon the method of breaking the conjugates. The second is that neither estradiol nor estrone levels show a continuous increase throughout pregnancy but peak in levels after midpregnancy, possibly indicating a different estrogen metabolite predominates in late pregnancy or that changes in conjugation (glucuronides and sulfates) may occur. Estrogens in pregnancy have been studied in the common marmoset by several investigators. The majority of urinary estrogens are excreted in the conjugated form during pregnancy in both marmosets and tamarins (23). With methods that remove simple conjugates (glucuronides and sulfides), the predominant estrogen is estradiol (23, 24, 25), although our preliminary results indicate that when using a method that removes double-conjugated estrogens (solvolysis) estrone is the predominant estrogen. Estriol is secreted in much lower levels than estradiol and estrone (23).
Estrogen precursors, DHEA and its sulfate, are produced by the fetal and maternal adrenals and converted into estrogens in the placenta. However, the proportion of DHEA that is produced by the fetus increases steadily during pregnancy (14). These large amounts of estrogens are excreted into the urine of the mother (26). Estrogens increase steadily throughout human pregnancy, but estriol shows the most pronounced steady increase. Although the largest concentration of urinary estrogen during pregnancy comes from estriol, the measurement of total estrogens in pregnancy urine appears to be more reflective of the fetal/placental unit. Placental estrogens are secreted into the maternal circulation in unconjugated form (27). Additionally, unconjugated estrogens have a shorter half-life and therefore provide a more reliable method of testing fetal-placental estrogens. The combination of low levels of placental lactogen and urinary estrogen has been found to be the best indicator of placental insufficiency associated with retarded intrauterine growth in humans (4). In terms of differentiating pregnancy outcomes by estrogen concentrations in the marmoset, only excreted free estradiol was significantly different in energy-restricted pregnancies than in normal pregnancies proceeding to term, suggesting that this estrogen excretion product is of most value in differentiating likely pregnancy outcomes.
Future studies in energy-restricted pregnant marmosets will be necessary to delineate the mechanisms behind the restriction-induced pregnancy loss, but these results point to impaired placental formation and function as the likely basis of fetal demise and loss and to excreted CG and free estradiol as good markers of impaired function.
| Footnotes |
|---|
First Published Online October 13, 2004
Abbreviations: CG, Chorionic gonadotropin; DHEA, dehydroepiandrosterone.
Received June 5, 2004.
Accepted September 14, 2004.
| References |
|---|
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |