| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Unit, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Maternal Physiology and...III. Fetal-Placental Physiology...IV. Maternal Physiology and...V. Neonatal Physiology and...VI. Discussion and ConclusionsReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Maternal Physiology and...III. Fetal-Placental Physiology...IV. Maternal Physiology and...V. Neonatal Physiology and...VI. Discussion and ConclusionsReferences |
|---|
Although Albright and Reifenstein’s theory proved to be incorrect, it is now evident that mineralization of the fetal skeleton and continued skeletal growth in the infant both mandate a series of hormone-mediated adjustments in maternal calcium metabolism during pregnancy and lactation, respectively. These hormone-mediated adjustments normally satisfy the daily calcium needs of the fetus and infant without long-term consequences to the maternal skeleton. In addition, both fetal and neonatal calcium and bone metabolism are uniquely adapted to meet the specific needs of these developmental periods. The fetus must actively transport sufficient calcium across the placenta to meet the large demands of the rapidly mineralizing skeleton, whereas the neonate must quickly adjust to loss of placental calcium transport, while continuing to undergo rapid skeletal growth.
Here we review our present understanding of normal human calcium and bone metabolism during pregnancy, lactation, fetal development, and the neonatal period. We shall also discuss the relevant pathophysiology and management of clinical disorders of calcium and bone metabolism that can occur during these periods. Generally these conditions are due to preexisting disease (e.g., hyperparathyroidism) that is compounded by the alterations in calcium and bone metabolism naturally occurring during these reproductive periods.
Although the focus of this review is on human calcium physiology and pathophysiology, the animal literature will be closely considered as well. Many of our models for explaining human physiology are based on these animal studies, particularly since ethical constraints generally prevent all but observational studies from being performed during human pregnancy and fetal development. Where both human and animal data are available, we will point out several significant differences that have been found between the animal and human data. These differences illustrate the difficulty of extrapolating from the animal models in the absence of human data.
The literature reviewed in this paper was obtained from computerized searches of the MEDLINE database, manual searches of Index Medicus before 1966, and the bibliographies of individual articles and texts.
|
II. Maternal Physiology and Pathophysiology During Pregnancy |
|---|
|
TopAbstractI. IntroductionII. Maternal Physiology and...III. Fetal-Placental Physiology...IV. Maternal Physiology and...V. Neonatal Physiology and...VI. Discussion and ConclusionsReferences |
|---|
The pregnant rat has typically been used as a model for studying
calcium metabolism during pregnancy, but the adaptive strategies of the
rat differ importantly from those of the human (Table 1
). These differences probably reflect
the large litter size (six to 12 fetuses) and the short gestational
period (22 days) of the rat; the rat must deliver 12 mg of calcium per
fetus between day 17 of gestation and term (6).
|
.
|
In contrast, the serum total and ionized calcium have been reported to fall during the last several days of pregnancy in the rat (21). Maternal losses of calcium to a litter of rapidly growing fetuses may exceed the maternal capacity to maintain a normal serum calcium level. Indeed, larger litter sizes correlated with lower serum calcium in pregnant rats (22). In white-tailed deer, the corrected serum calcium falls in the last 1 to 2 weeks of gestation (23). Pregnant ewes have a mild decrease in total serum calcium over the last 6 weeks of pregnancy, likely due to the fall in serum albumin (24); moreover, in one study, about 13% of Awassi fat-tail ewes were found to develop signs and biochemical evidence of hypocalcemia in the last month of pregnancy (25). Therefore, data from several animal models suggest that maternal blood calcium regulation may be disrupted by fetal demands in late pregnancy.
2. Phosphate. Serum phosphate levels are normal throughout pregnancy in humans and animals, as is the renal tubular reabsorption of phosphate (14, 17, 26, 27, 28, 29).
3. PTH. The bulk of published human data on PTH levels in pregnancy was obtained from studies that used early-generation PTH RIAs (18, 26, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40); some of the more frequently cited studies reported high maternal serum levels of PTH in the latter half of pregnancy (18, 30, 31, 32, 33, 34, 39). These data must now be reinterpreted, because it is now known that these PTH RIAs were insensitive and heterogeneously measured multiple different fragments of PTH, most of which were biologically inactive (41, 42).
With the advent of sensitive two-site immunoradiometric (IRMA) PTH assays that accurately determine the level of intact PTH (42), PTH levels have been typically found to be low-normal in the serum of pregnant women in all three trimesters (11, 12, 17, 19, 20, 27, 43). Five prospective, longitudinal studies found that the mean PTH level was in the low-normal range (i.e., <50% of the mean nonpregnant value) during the first trimester but increased steadily to the mid-normal range by the end of pregnancy (14, 15, 16, 27, 44). These findings have been independently validated by reports of normal nephrogenous cAMP levels (12, 26, 44) and low to normal PTH-like bioactivity (11) throughout human pregnancy (although this may be confounded by synthesis of nephrogenous cAMP due to the effects of PTHrP). Studies in primates suggest that the parathyroid glands may have less secretory reserve as pregnancy progresses; the incremental PTH response to acute EDTA-induced hypocalcemia in rhesus monkeys decreased across the trimesters (39).
In contrast to humans, rats develop secondary hyperparathyroidism late
in pregnancy. Normally, in late pregnancy, both maternal levels of
intact immunoreactive (45) and bioactive (46) PTH rise to exceed the
normal range, and the maternal ionized and total calcium levels decline
slightly (21). The parathyroid gland volume has also been reported to
increase during normal rat pregnancy (47, 48). In vitro
studies in pregnant rats indicate that the parathyroids secrete more
PTH at a given extracellular calcium concentration, when compared with
parathyroid cells taken from nonpregnant rats (49). The PTH levels
begin to rise earlier in gestation, and peak at higher levels, in
pregnant rats fed a modestly calcium-restricted diet (21, 50). This
increase in PTH during late pregnancy is critical for normal maternal
calcium homeostasis; parathyroidectomized pregnant rats can exhibit
signs of tetany in the last 2–4 days of gestation and death during the
birthing process (51, 52, 53, 54). In parathyroidectomized pregnant rats,
dietary intake and weight gain decline, while serum
1,25-dihydroxyvitamin D and intestinal calbindin9K-D levels
fall (52, 53, 54). Maternal tetany coincides with the time onset of rapid
fetal accretion of calcium (6); therefore, the parathyroidectomized
pregnant rat has compromised dietary intake and intestinal calcium
absorption at the time of peak fetal demand for calcium. The calcium
abnormalities can be completely prevented when the rats are fed a
high-calcium, low-phosphorus diet. Taken together, these observations
indicate that rats (but not humans) normally develop a form of
secondary hyperparathyroidism during late pregnancy in response to the
fall in the maternal serum calcium level. Rats may be more dependent on
PTH-mediated bone resorption and PTH-induced 1
-hydroxylase
up-regulation during late pregnancy, at a time when the combined
calcium need of a litter of fetuses is at its peak.
In summary, immunoreactive and bioactive PTH levels are in the low-normal range during early human pregnancy and are in the mid-normal range at term; in contrast, immunoreactive and bioactive PTH levels in rats are normal in early pregnancy but exceed the normal range in late gestation.
4. 1,25-Dihydroxyvitamin D. Cross-sectional studies have found
that the serum level of 1,25-dihydroxyvitamin D more than doubles early
in the first trimester in human pregnancy (12, 27, 36, 37, 55, 56, 57, 58).
Longitudinal studies have found that the levels of both free and bound
1,25-dihydroxyvitamin D are doubled, and that this increase is
maintained until term (15, 20, 41, 57, 58, 59). Although clearance of
1,25-dihydroxyvitamin D has not been studied during human pregnancy, in
pregnant rats, sheep, and rabbits the increased 1,25-dihydroxyvitamin D
levels were due to increased production, and not decreased metabolic
clearance, of 1,25-dihydroxyvitamin D (60, 61, 62, 63). In vitro
measurements in homogenates of maternal kidney from rabbits and guinea
pigs show that the renal 1-hydroxylase may be up-regulated 2- to
5-fold (64, 65). The increase in the 1,25-dihydroxyvitamin D level
begins while the PTH level is in the low-normal range in humans
(Section II.B.3, above); this may indicate that PTH does
not mediate the up-regulation of the maternal renal
1-hydroxylase during early human pregnancy. Furthermore,
parathyroidectomy in pregnant sheep reduces, but does not eliminate,
the pregnancy-related increase in 1,25-dihydroxyvitamin D (66). Other
potential direct or indirect regulators of the 1-hydroxylase include
PTHrP (Section II.B.6, below), estradiol, PRL, and
placental lactogen. Estradiol (67), PRL (68, 69), and placental
lactogen (69) acutely stimulate the 1-hydroxylase in
vitro, and placental lactogen (but not PRL) raised the serum
1,25-dihydroxyvitamin D levels in hypophysectomized, nonpregnant rats
(70). The effect of estradiol on the 1-hydroxylase has been
confirmed in vivo by the observation that estrogen
replacement in postmenopausal women increases the free and total serum
1,25-dihydroxyvitamin D level (71). However, an effect of PRL in
vivo has not been confirmed, since hyperprolactinemic patients
showed no alteration in 1,25-dihydroxyvitamin D levels (72). Further,
in pregnant women, the high 1,25-dihydroxyvitamin D levels of pregnancy
did not correlate with serum PRL, estrogens, or human placental
lactogen (73).
In addition to the renal 1-hydroxylase, 1-hydroxylase activity
found in maternal decidua, placenta, and fetal kidneys may also add
1,25-dihydroxyvitamin D to the maternal circulation during pregnancy
(59, 74, 75, 76, 77). To test this hypothesis,
[3H]25-hydroxyvitamin D was administered to pregnant rats
after bilateral maternal nephrectomy (74, 78). Newly synthesized
(i.e., tritiated) 1,25-dihydroxyvitamin D appeared in
the maternal circulation of nephrectomized pregnant rats (but not in
nonpregnant nephrectomized rats). Although this study indicates that
extrarenally produced 1,25-dihydroxyvitamin D can reach the maternal
circulation (74), the specific extrarenal sites and the amounts of
1,25-dihydroxyvitamin D produced could not be ascertained. Data from
the Hannover pig model (autosomal recessive 1-hydroxylase
deficiency) indicate that the amounts contributed by these extrarenal
sites may be insignificant. In pregnant sows homozygous for absence of
1-hydroxylase activity, serum levels of 1,25-dihydroxyvitamin D were
very low, comparable to the nonpregnant values (79). The presence of
heterozygous fetuses did not increase the circulating level of
1,25-dihydroxyvitamin D in the homozygous sows (79). The same gene
controls renal and decidual 1-hydroxylase activity in this model
(77). A single case report of a human patient on chronic hemodialysis
found 1,25-dihydroxyvitamin D levels of 10–15 pg/ml during pregnancy;
these levels were higher than the nonpregnant level in the same
patient, but were far lower than in normal pregnancy (80). It is,
therefore, likely that increased maternal production of
1,25-dihydroxyvitamin D is mainly due to increased activity of
maternal, renal 1-hydroxylase and not to large contributions from
extrarenal sites.
Again, the pregnant rat model differs somewhat from the human, in that
the maternal rise in 1,25-dihydroxyvitamin D level does not occur in
rats until the time of fetal skeletal mineralization in late gestation
(22, 45, 45, 81, 82), at which time the serum PTH levels rise above
normal (22, 45) and serum ionized calcium levels fall (21, 22). Larger
litter sizes correlate with higher maternal 1,25-dihydroxyvitamin D
levels (22). These studies suggest that the effect of PTH on the renal
1-hydroxylase may dominate the production of 1,25-dihydroxyvitamin D
during late pregnancy in the rat.
Serum 25-hydroxyvitamin D levels are unchanged in human pregnancy, and 24,25-dihydroxyvitamin D levels are lower in pregnant women than in controls (35). Supplementation with 1000 IU of vitamin D3 daily after the first trimester in humans did not affect maternal calcium, phosphate, PTH, and 1,25-dihydroxyvitamin D levels; this suggests that the changes in calcitropic hormone levels observed in human pregnancy are not the result of occult vitamin D deficiency (83). Maternal vitamin D deficiency in the rat has been associated with reduced fertility and smaller litter sizes, and up to 20% of pregnant, vitamin D-deficient rats may die of hypocalcemia near term (84, 85).
In summary, free and total 1,25-dihydroxyvitamin D levels rise early in
human pregnancy to peak at twice the normal range, while in rats the
1,25-dihydroxyvitamin D level does not rise until late gestation. These
increases appear to be due to increased production of
1,25-dihydroxyvitamin D by the maternal kidneys, with possibly small
contributions from maternal decidua, placenta, and fetal kidneys. PTH
may be less important during pregnancy in humans compared with rats in
mediating this rise in 1-hydroxylase activity.
5. Calcitonin. Serum calcitonin levels in human pregnancy have generally been reported to be higher than nonpregnant values, with at least 20% of values exceeding the normal range (14, 34, 36, 37, 86, 87, 88, 89). Several human studies reported that calcitonin levels were not elevated in pregnancy (15, 18, 33, 35); however, these studies were flawed by the use of improper controls. For example, in some of these studies, postpartum measurements in the same women were used as the baseline, and it has since been shown that calcitonin is also elevated in the postpartum period (see Section IV.B.5, below). Similar data from monkeys (39), sheep (77, 90, 91), deer (23), goats (77), and rats (92) have confirmed that the maternal calcitonin level is elevated during pregnancy. No clearance data are available for humans or other animals, but the increased level of calcitonin is generally thought to reflect increased synthesis.
Thyroidal C cells, breast, and placenta are sites of calcitonin synthesis during pregnancy (93, 94). It is not surprising, therefore, that a rise in calcitonin is found in totally thyroidectomized women, most likely due to calcitonin synthesized by the placenta and breast (93, 94). In pregnant rhesus monkeys, acute calcium infusions led to a progressively greater calcitonin response across the trimesters, which may indicate greater secretory reserve of the thyroidal C cells and placenta (39).
It has been speculated that elevated calcitonin protects the maternal skeleton from excessive resorption of calcium, a hypothesis that has been difficult to prove. Indeed, the physiological role of calcitonin in human calcium and skeletal metabolism has not been established (95). No adequate model of experimental calcitonin deficiency has been created, partly because the extrathyroidal sites of calcitonin synthesis were not appreciated at the time. All models used total thyroidectomy with parathyroid gland autotransplantation and thyroid hormone replacement in pregnant goats or rats (53, 91, 96, 97, 98). In none of these models was the serum calcitonin or TSH measured to determine whether a calcitonin-deficient, euthyroid state had been attained. Thus, although these models suggested that an intact thyroid gland protected the maternal skeleton from loss of bone mineral during pregnancy, these findings remain to be confirmed by more rigorously controlled models.
In summary, calcitonin levels are increased during pregnancy in humans and animals, partly due to extrathyroidal synthesis in the placenta and breast. The possible role of calcitonin in protecting the maternal skeleton from increased resorption during pregnancy needs more study.
6. PTH-related protein (PTHrP). PTHrP was originally identified in 1987 as the cause of humoral hypercalcemia of malignancy (99). PTHrP has been postulated to be a prohormone, which is processed into several different circulating fragments or hormones, each of which, in turn, may have different functional roles and specific receptors (100). PTHrP has partial homology in its first 13 amino acids to PTH (101, 102, 103) and activates the common PTH/PTHrP receptor (42). Amino-terminal forms of PTHrP (PTHrP 1–34, 1–86, or 1–141) resemble PTH in their actions on kidney and bone (104) and can inhibit acetylcholine-induced uterine contractions in the rat (105). Levels of PTHrP decreased acutely in the amnion and myometrium at the time of onset of labor in humans (106). It has been suggested that amino-terminal forms of PTHrP may, therefore, have a role in regulating the onset of labor (106). A midmolecular form of PTHrP stimulates placental calcium transport in the fetus (Section III.C, below), although its possible role in the mother is unclear. The carboxyl-terminal portion of PTHrP, termed "osteostatin," is able to inhibit osteoclastic bone resorption in some in vitro assays (107, 108) and in rats in vivo (109); therefore, this fragment of PTHrP could have a role in protecting the maternal skeleton during pregnancy.
The development of RIAs for PTHrP has concentrated on detecting the PTH-like amino-terminal fragments of PTHrP and has thus far largely ignored the detection of other fragments that might be biologically active. Therefore, no data are available on the levels of midmolecular or carboxyl-terminal fragments of PTHrP during pregnancy compared with controls. An early RIA that used an antibody to PTHrP 1–34 found no elevation of PTHrP in pregnancy (110). Newer, more sensitive two-site immunoradiometric assays that measure forms of PTHrP that encompass amino acids 1 through 86 have found a significant increase in the maternal PTHrP level, beginning as early as weeks 3 to 13 of human pregnancy (44, 111). This increase is not due to any change in the clearance of PTHrP 1–34, 1–86, or 1–141 during pregnancy, as determined in sheep (112, 113). The increase in amino-terminal PTHrP, by activating the PTH/PTHrP receptor in kidney and bone, may well explain (at least in part) the increase in 1,25-dihydroxyvitamin D and ionized calcium, and the decrease in PTH levels, found during human pregnancy.
The source of PTHrP in the maternal circulation during pregnancy is not established, but several candidate sites are known. PTHrP is produced by the placenta (114), amnion (106), decidua (106), umbilical cord (115), and fetal parathyroid glands (116) and potentially might reach the maternal circulation. PTHrP produced by the breast tissue is detectable in human colostrum (117), and it is produced as early as day 14 of pregnancy by the mammary glands of the rat (118).
Overproduction of PTHrP by the breast might explain the development of hypercalcemia at 24 weeks of gestation in a woman with massive (4.5 kg) mammary hyperplasia of pregnancy, associated with hypercalciuria, hypophosphatemia, and undetectable PTH levels (119). Bilateral mastectomies in the second trimester of that same pregnancy corrected the hypercalcemia and the suppressed PTH level (119).
In summary, PTHrP may be made available to the maternal circulation by several different maternal and fetal sources. PTHrP fragments encompassing amino acids 1–86 are increased in the maternal circulation during pregnancy and may contribute to the elevations in 1,25-dihydroxyvitamin D and blood calcium, and suppression of PTH, noted during pregnancy. The true quantitative importance of PTHrP in maternal physiology needs to be established.
7. Other hormones. Pregnancy induces a dramatic rise in other hormones, including the sex steroids, PRL, and placental lactogen. The possibility that each of these, in turn, may have direct or indirect effects on calcium and bone metabolism during pregnancy has been largely unexplored. There is some evidence to suggest that PRL and placental lactogen may increase the intestinal transport of calcium (70, 120, 121), reduce urinary calcium excretion (122, 123), and stimulate synthesis of PTHrP (124) and 1,25-dihydroxyvitamin D (68, 69). This is discussed in more detail in the relevant sections.
C. Intestinal absorption of calcium
Calcium is absorbed throughout the small intestine, a small
portion by active transport in the duodenum and proximal jejunum, and
the major portion by passive mechanisms in the distal jejunum and ileum
(125). Mineral balance and calcium kinetic studies in humans using
stable isotopes of calcium (48Ca, 44Ca,
42Ca) have consistently found a positive calcium balance
and a doubling of the intestinal absorption of calcium during human
pregnancy from as early as 12 weeks of gestation (the earliest time
point studied) (27, 126, 127). By studying the effect of an oral
calcium load on serum calcium and urine calcium excretion, other
investigators indirectly confirmed that intestinal calcium absorption
must be increased in all trimesters (12, 128). The results of these
studies led to speculation that the increase was mediated by
1,25-dihydroxyvitamin D, and this appeared to be confirmed when
elevated levels of 1,25-dihydroxyvitamin D were found during human
pregnancy (Section II.B.4, above). 1,25-Dihydroxyvitamin D
probably stimulates intestinal calcium absorption by increasing the
synthesis of proteins, including the intestinal vitamin D-dependent
calcium-binding protein, calbindin9K-D. Protein and mRNA
levels of calbindin9K-D increase in the intestines of mice
and rats during pregnancy and plateau when both maternal
1,25-dihydroxyvitamin D levels and the efficiency of intestinal calcium
absorption are at peak levels (129, 130, 131). Maternal vitamin D deficiency
in rodents reduces the rise in the intestinal expression of
calbindin9K-D (132, 133), while 1,25-dihydroxyvitamin D
administration can restore it (133).
The rise in intestinal absorption of calcium occurs by midpregnancy in rats, before the onset of rapid skeletal mineralization in the fetus (45). The doubling of intestinal absorption persists in parathyroidectomized rats (134) and may, therefore, be independent of PTH regulation. The early increase in intestinal calcium absorption may allow the pregnant mother to accrete calcium (probably in the maternal skeleton), before the peak fetal demand for calcium in late pregnancy. Consistent with this hypothesis, it has been estimated from isotope studies in the pregnant rat that 92% of fetal skeletal calcium content was absorbed from the maternal diet at some point during pregnancy (135). Further, several investigators have found that pregnant rats normally store calcium during the first half of pregnancy (136), such that by the end of pregnancy, the calcium content of the femurs is unchanged (137). Inadequate accretion of calcium early in pregnancy may lead to a net loss of maternal skeletal calcium later in pregnancy. For example, under dietary calcium restriction, pregnant rats (138, 139) and goats (98) have reduced calcium content in their long bones by the end of gestation. Similarly, maternal vitamin D deficiency has been found to cause maternal skeletal demineralization by the end of pregnancy (140).
PRL treatment of pregnant, vitamin D-deficient rats resulted in an increase in the intestinal absorption of calcium; PRL might, therefore, have an effect on the intestine independent of 1,25-dihydroxyvitamin D (120). This is further supported by studies in everted gut sacs of nonpregnant, hypophysectomized rats, where PRL and placental lactogen stimulated the intestinal transport of calcium (70, 121). Also in rats, the increase in duodenal calcium absorption has been found to precede the rise in the 1,25-dihydroxyvitamin D level by 1 week, suggesting that the intestinal effect is not dependent solely on vitamin D (45, 141). Even in the absence of vitamin D, pregnancy in rats is associated with hypertrophy of the small intestine and a doubling of intestinal absorption of calcium (141, 142). Furthermore, rats hypocalcemic from vitamin D deficiency developed a progressive rise in serum calcium levels during pregnancy, despite unchanged serum PTH levels (143). However, an independent effect of PRL on intestinal calcium absorption could not be demonstrated in studies on humans. Hyperprolactinemic patients showed no alteration in the intestinal absorption of calcium (72).
In summary, intestinal calcium absorption is increased 2-fold early in human and rat pregnancy, probably through a 1,25-dihydroxyvitamin D-mediated increase in intestinal calbindin9K-D and other proteins. PRL and placental lactogen (or possibly other factors) may mediate part of the normal pregnancy-related increase in intestinal calcium absorption. The early rise in intestinal calcium absorption may allow the maternal skeleton to store calcium in advance of the peak fetal demands later in pregnancy. The increased intestinal calcium absorption appears to be a major maternal adaptation to meet the fetal need for calcium.
D. Renal handling of calcium
Pregnancy is associated with an increase in creatinine clearance
and glomerular filtration rate (144, 145). The 24-h urine calcium
excretion is increased as early as the 12th week of gestation (the
earliest time point studied), and averages 300 ± 61 mg in the
third trimester (levels in the hypercalciuric range are not uncommon)
(12, 14, 20, 27, 146, 147). Since fasting urine calcium values are
normal or low, the increase in 24-h urine calcium reflects the
increased intestinal absorption of calcium (absorptive hypercalciuria)
(12, 28, 44). A similar 2-fold increase in urinary calcium excretion
has been observed in the pregnant rat from the second week of gestation
(148). Although PRL and placental lactogen have been shown to reduce
urinary calcium excretion in nonpregnant rabbits in vivo
(122, 123), the effect (if any) of either hormone on the kidneys of
pregnant humans and rats must be very modest.
Interestingly, preeclampsia and pregnancy-induced hypertension (PIH) have been associated with hypocalciuria (147, 149, 150, 151, 152). Further studies have found the hypocalciuria to be associated with low 1,25-dihydroxyvitamin D levels (149, 150, 151, 152), but to be independent of PTH, calcitonin, or ionized calcium levels (147, 149, 150, 151). The finding of hypocalciuria prompted a large trial of calcium supplementation in pregnant women, which recently reported no benefit in preventing preeclampsia or PIH (153). These abnormalities in 1,25-dihydroxyvitamin D and urine calcium excretion are, therefore, probably secondary to a primary renal tubular defect occurring in preeclampsia and PIH and are likely not the primary cause of the hypertension (149).
E. Skeletal calcium metabolism
1. Bone formation and resorption. Histomorphometric parameters
of both bone formation and osteoclast-mediated resorption are increased
during pregnancy in rats (154). Pregnant beagle dogs also show
histomorphometric evidence of increased bone turnover in iliac
trabecular bone (155). Despite evidence of increased turnover, bone
mineral content during pregnancy in rats does not change (137, 140, 154, 156). In contrast, pregnant ewes have a 20% decrease in skeletal
calcium content during gestation (157).
Comparable histomorphometric data are not available for human pregnancy, but markers of bone formation and resorption have been assessed. Generally speaking, such indices are more reliable for measuring changes in bone resorption than bone formation (158, 159). Several markers of bone resorption (tartrate-resistant acid phosphatase, deoxypyridinoline/creatinine, pyridinoline/creatinine, and hydroxyproline/creatinine) are low in the first trimester but rise steadily to peak at values up to twice normal in the last trimester (27, 44, 160, 161). In contrast, osteocalcin, a marker of bone formation, is low or undetectable early in gestation and sometimes rises to normal levels by term (15, 27, 161, 162, 163). Other markers of bone formation (procollagen I carboxypeptides, bone-specific alkaline phosphatase) are low in the first trimester and have been found to remain low (44) or rise to normal or above in the last trimester (27, 160). Total alkaline phosphatase rises early in pregnancy due to contributions from the placental fraction, and, therefore, is not a useful marker of bone formation in pregnancy (14, 44).
Taken together, the histomorphometric data from animals, and the changes in the markers of bone formation and resorption in humans, indicate that bone turnover is probably low in the first half of pregnancy, but may be increased in the third trimester. The third trimester increase in bone turnover corresponds to time of the peak rate of calcium transfer to the fetus and may result from mobilization of skeletal calcium stores (which contain 99% of the body’s stores of calcium) to help supply the fetus.
2. Bone density. Concerns about fetal radiation exposure have resulted in few studies of changes in maternal bone mass during pregnancy; these studies used techniques that are far less precise or reproducible than the current standard, dual x-ray absorptiometry (DXA) (164, 165). Of the scant data available, an early study used x-ray spectrophotometry of the radius and femur to demonstrate a progressive decrease in trabecular bone density during pregnancy (166). Using more modern techniques, four prospective studies of bone density during pregnancy did not find a significant change in cortical or trabecular bone density, as respectively determined by single photon absorptiometry (SPA) and/or dual-photon absorptiometry (DPA) (28, 146, 167, 168). Another study found a significant decrease in bone mineral density of the femoral neck and radial shaft, but no change in lumbar bone density, by comparing preconception SPA and DPA measurements to those taken 6 weeks postpartum (169). Most recently, cross-sectional (170) and longitudinal studies (161, 171) have found a progressive decrease during pregnancy in indices thought to correlate with bone mineral density, as determined by ultrasonographic measurements of the os calcis in all three trimesters.
The majority of retrospective, epidemiological studies of pre- and postmenopausal women have found no association of parity with bone density or fracture risk (172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191). In contrast, several other studies found increased parity to be beneficial, as indicated by a slightly greater lumbar (192, 193), femoral (194), or radial bone density (194, 195, 196) and decreased hip fracture risk (197, 198). Four remaining studies linked parity to somewhat decreased lumbar bone density (199, 200, 201) or increased hip fracture risk (202). An epidemiological study of healthy women aged 21 to 95 found divergent effects of parity at different anatomical sites. Femoral neck bone mineral density was decreased in parous women by 1.5% per live birth, while lumbar spine bone density was not influenced by parity (203). Among the studies that found no significant association of parity, several reported that a first pregnancy as an adolescent was associated with decreased bone density (178, 190, 195), possibly because the fetal calcium demands of pregnancy reduce the peak bone mass that is eventually achieved in the adolescent. Overall, many of these epidemiological studies had significant methodological limitations, specifically the difficult problem of retrospectively separating out the effects of parity from those of lactation. Nevertheless, it may be reasonable to conclude from these studies that if parity has either a positive or a negative effect on bone density or fracture risk, it must be only a very modest effect.
Therefore, although changes in serum and urine markers of bone formation and resorption have indicated that bone turnover may be increased in the third trimester, it is impossible to determine from the available bone density data whether there is any acute change in bone mineral during human pregnancy. Further, it is also unknown whether any such acute change has a long-term effect on the calcium content or fragility of the maternal skeleton.
3. Osteoporosis in pregnancy. The rare presentation of idiopathic osteoporosis in a woman of child-bearing age has often been associated with a recent pregnancy, as noted by Albright and Reifenstein (1) and other early case reports (204, 205, 206). The exact prevalence of the condition is uncertain. The theory that pregnancy might cause osteoporosis (as proposed by Albright and Reifenstein) was disputed by an early observational study of five women with symptomatic, severe osteoporosis presenting in a first pregnancy (207). In subsequent pregnancies, these women were found to have no worsening of their condition, but the parameters used (new pain or fracture, worsening of osteopenia on plain roentgenograms) were crude and insensitive by methods available today (207). Despite better documentation of the absence of known causes of decreased bone density in more recent case reports (208, 209, 210, 211, 212), it has not been possible to exclude the possibility that low peak bone mass and/or an accelerated bone resorptive state preceded the pregnancy and simply became clinically obvious in pregnancy. In addition, some reported cases of osteoporosis in pregnancy have been clearly confounded by the presence of other recognizable causes of secondary osteoporosis, such as chronic heparin, anticonvulsant, or corticosteroid therapy (209, 212). In two documented cases of osteoporosis diagnosed in pregnancy, the female progeny were found at age 10 to also have low bone mineral density (213). This finding suggested that a shared genetic or environmental factor (and not pregnancy) was the cause of osteoporosis in the mothers and daughters. The limited data from bone biopsy typically show no evidence of osteomalacia, but only mild osteoporosis or normal architecture (208, 212). It remains intriguing to speculate that some of these rare cases of osteoporosis presenting in pregnancy may result from excessive resorption of calcium from the maternal skeleton, perhaps in the setting of inadequate intake of calcium, low stores of 25-hydroxyvitamin D, or an excessive rise in PTHrP in the maternal circulation (see also the discussion of osteoporosis in lactation, Section IV.E.3, below). Nevertheless, these rare cases may simply represent idiopathic osteoporosis occurring in pregnant women by mere chance.
A second (also rare) form of pregnancy-associated osteoporosis is a focal, transient osteoporosis of the hip (214, 215, 216, 217). Typically these patients present with unilateral or bilateral hip pain, limp, and/or hip fracture in the third trimester (214, 216, 218, 219). Radiolucency of the femoral head and neck was recognized on plain radiographs taken in early reports of this condition (215, 220); and more recently, DXA measurements have shown that the bone density of the symptomatic femoral head and neck is reduced (218). Magnetic resonance imaging (MRI) of the affected femoral head in one patient showed a joint effusion and images suggesting increased water content of the femoral head and marrow cavity (221). Routine serum chemistries are typically normal (222). Alkaline phosphatase and urine hydroxyproline have been reported to be elevated (215, 218, 223); however, the interpretation of these findings is uncertain, since control measurements from normal pregnant women were not compared. Intriguingly, the decreased bone mineral density of the femoral head and neck typically resolves within 2 to 6 months postpartum (214, 218, 219), including the MRI findings (221). Patients generally require only pain relief and continued mobilization for this self-limited condition. The fact that this rare condition is typically localized to one or both femoral heads, and not the rest of the skeleton, suggests that it is not the result of a generalized increase in skeletal resorption. Several theories have been proposed to explain this condition, including femoral venous stasis due to pressure from the pregnant uterus, a form of Sudeck’s atrophy, or reflex sympathetic dystrophy (causalgia), ischemia, trauma, viral infections, marrow hypertrophy, immobilization, and fetal pressure on the obturator nerve (214, 215, 216, 220). As yet, the etiology of transient osteoporosis of the hip in pregnancy remains unclear; its association with pregnancy may be not causal but incidental. In any case, it appears likely that this disorder is not a manifestation of altered calcitropic hormone levels or mineral balance during pregnancy.
F. Primary hyperparathyroidism
The presentation of primary hyperparathyroidism in pregnancy
raises important diagnostic and management considerations. Many cases
are asymptomatic, detected by routine prenatal biochemical tests or
after the presentation of hypocalcemia in the neonate. Several normal
pregnancy-related changes in calcium and PTH physiology (noted above)
may obscure the diagnosis of mild primary hyperparathyroidism. These
include the fall in total serum calcium, the rise in the corrected
serum calcium (111, 224), the fall in the intact PTH level (14, 15, 16, 44), and the rise in the 24-h urinary excretion of calcium, often into
the hypercalciuric range (14, 27) (see also Sections II.B and
II.D, above).
Although maternal primary hyperparathyroidism in pregnancy is probably a rare condition (there are no data available on its prevalence), it has been associated in the literature with an alarming rate of adverse outcomes in the fetus, including a 30% rate of spontaneous abortion or stillbirth (225, 226). In the neonatal period, a 50% rate of tetany and a 25% rate of neonatal death has been reported (225, 227). These adverse outcomes are thought to result from suppression of the fetal parathyroid glands; this suppression may occasionally be prolonged for months (228, 229). PTH cannot cross the placenta (230, 231, 232); therefore, the fetal parathyroid suppression is thought to result from increased net calcium flux across the placenta to the fetus, facilitated by maternal hypercalcemia. Evidence from animal models has confirmed that acute elevations in maternal serum calcium cause an increase in fetal serum calcium, and a fall in fetal PTH level (233). However, whether chronic maternal hypercalcemia has the same effect on fetal serum calcium, or placental calcium transport, has not been determined.
Surgical correction of primary hyperparathyroidism during the second trimester, to prevent fetal and neonatal complications, has been almost universally recommended (226, 234, 235, 236). Several case series have found elective surgery to be well tolerated and to dramatically reduce the rate of adverse events when compared with the earlier cases reported in the literature (234, 235, 237, 238). However, many of the women in those early cases were symptomatic and had nephrocalcinosis or renal insufficiency. Those early case reports may also have reflected reporting bias of adverse fetal and neonatal outcomes. Whether the milder, asymptomatic form of primary hyperparathyroidism commonly seen today has the same risk of adverse fetal or neonatal outcomes has not been determined. In several case reports, mild elevations in maternal serum calcium were followed without operative intervention, and no adverse fetal or neonatal outcome occurred (239, 240). However, in other cases the mild hypercalcemia of both asymptomatic primary hyperparathyroidism and familial hypocalciuric hypercalcemia has been reported to cause neonatal parathyroid suppression and tetany (241, 242, 243). Nevertheless, it is probably reasonable to follow cases of asymptomatic primary hyperparathyroidism with mild hypercalcemia conservatively and to reserve surgery in the second trimester for patients that are symptomatic or have more severe hypercalcemia. If surgery is deferred, the neonate must be monitored closely for the development of hypocalcemia.
G. Hypoparathyroidism and pseudohypoparathyroidism
As described earlier (Section II.B.4), free and bound
1,25-dihydroxyvitamin D levels normally double during human pregnancy
in the presence of low-to-normal intact PTH levels, and, therefore, it
is likely that PTH does not mediate the pregnancy-related rise in
1,25-dihydroxyvitamin D production. Other hormones of pregnancy, such
as estrogen, PTHrP, and perhaps placental lactogen and PRL, may
regulate the increased production of 1,25-dihydroxyvitamin D by
maternal kidney and decidua. Also, placenta and fetus may contribute to
the maternal increase in 1,25-dihydroxyvitamin D.
In multiple case reports, pregnant hypoparathyroid women have been found to have fewer hypocalcemic symptoms, a rise in the serum calcium, and decreased dependence on supplemental calcitriol to maintain a normal serum calcium (244, 245, 246, 247, 248, 249, 250, 251, 252). This finding is consistent with a limited role for PTH in the pregnant woman and suggests that an increase in 1,25-dihydroxyvitamin D and/or increased intestinal calcium absorption will occur in the absence of PTH. The literature on hypoparathyroidism in pregnancy is not entirely consistent on this point, since in other case reports the calcitriol dosage was increased for a variety of reasons (some incompletely documented) (253, 254, 255, 256, 257). Despite these contrasting views on the natural history of hypoparathyroidism in pregnancy, there is general agreement (244, 245, 248, 253, 254, 258) that in late pregnancy and the puerperium, hypercalcemia may result unless the calcitriol is discontinued, or the dosage is decreased below the prepartum requirement. Since this effect is even more pronounced in those who breast-feed, and since PTHrP is found at high concentrations in the breast during late pregnancy and lactation (further discussed in Section IV.B.6, below), the pregnancy-related rise in 1,25-dihydroxyvitamin D production may be regulated by PTHrP (secreted from the breast) in these hypoparathyroid women.
Calcitriol (rather than vitamin D or calcifediol) has typically been prescribed for hypoparathyroidism in pregnancy, and the dosage needed may range from 0.5–3.0 µg daily. Chronic maternal hypocalcemia must be avoided because it has been associated with the development of intrauterine hyperparathyroidism and death in the fetus (Section III.F, below).
Further illumination of the role of PTH in pregnancy has come from cases of pseudohypoparathyroidism in pregnancy. Pseudohypoparathyroidism is a heterogeneous group of genetic syndromes characterized by hypocalcemia due to PTH resistance (259). Although the data are limited, Breslau and Zerwekh (260) noted a normalization of serum calcium levels in two pregnant women with pseudohypoparathyroidism (probably type 1b). Before pregnancy the patients had hypocalcemia, markedly elevated PTH levels, and low 1,25-dihydroxyvitamin D levels. During four pregnancies (two for each patient), the serum calcium levels were normal, their PTH levels were halved, and the 1,25-dihydroxyvitamin D levels increased 2- to 3-fold. Contributions of 1,25-dihydroxyvitamin D from placental and fetal sources might have accounted for these findings; Zerwekh and Breslau (261) noted elsewhere that the placental production of 1,25-dihydroxyvitamin D was no different between placentas obtained from pseudohypoparathyroid women and controls. Alternatively, it is possible that the hormonal milieu of pregnancy lessened the renal resistance to PTH and PTHrP and thereby increased the formation of 1,25-dihydroxyvitamin D. It is apparent from Breslau’s observations that estrogens alone cannot be the explanation for such an improvement during pregnancy, because the same two pseudohypoparathyroid women were not improved by treatment with an oral contraceptive. In any case, calcitriol supplementation in these patients should be monitored carefully and adjusted during pregnancy. The progeny of these pregnancies are also at risk of intrauterine, fetal hyperparathyroidism (262, 263), perhaps because of relative maternal hypocalcemia during pregnancy.
H. Summary
The fetal demand for calcium, which largely occurs during the
third trimester, is met by a doubling of free and bound maternal
1,25-dihydroxyvitamin D levels, which, in turn, partly mediate a
doubling of the intestinal absorption of calcium. Some of the increased
intestinal calcium absorption may be mediated by PRL or other hormones
of pregnancy. Further, the increase in 1,25-dihydroxyvitamin D may be
largely independent of changes in PTH, since PTH levels are typically
low or normal at the time of the increase in 1,25-dihydroxyvitamin D.
The increased calcium intake and absorption leads to a marked increase
in renal calcium excretion (absorptive hypercalciuria). The serum
ionized calcium is normal, despite a fall in total serum calcium caused
by a reduction in the albumin-bound fraction. Calcitonin and PTHrP are
both elevated, particularly in the latter half of gestation, but the
physiological importance of these hormones in pregnancy is not known.
The typical changes in calcium and calcitropic hormone levels during
pregnancy are depicted schematically in Fig. 1
.
Bone resorption is increased during late pregnancy, as evidenced by a rise in the levels of serum and urine markers of bone resorption in the third trimester, and this may indicate that maternal skeletal calcium stores are mobilized during the time of rapid fetal accretion of calcium. As noted at the beginning of Section II.E.2, bone density studies during pregnancy have been of insufficient precision to determine whether this increased bone resorption results in significant loss of skeletal calcium during pregnancy or the third trimester. Retrospective epidemiological studies (although not definitive) have generally found no effect of parity on the risk of osteoporosis or fractures in later life. Uncommonly, pregnancy may be associated with osteoporosis and fractures, particularly if the woman enters pregnancy with a low peak bone mass. A distinct disorder, focal, transient osteoporosis of the hip in pregnancy, is not likely due to altered calcitropic hormone levels and calcium physiology.
Primary hyperparathyroidism in pregnancy has been classically associated with adverse fetal or neonatal outcomes, but the milder, asymptomatic form of primary hyperparathyroidism most often seen today may not share such outcomes. Maternal hypoparathyroidism may be improved in pregnancy by increased intestinal absorption of calcium, possibly mediated by increased production of 1,25-dihydroxyvitamin D caused by PTHrP or some other non-PTH factor. A similar improvement in biochemical indices has been seen in pregnant women with pseudohypoparathyroidism. In both hypoparathyroid and pseudohypoparathyroid women, maternal hypocalcemia may adversely affect the fetus and must be avoided.
The pregnant rat model differs from the human condition in several
important respects (Table 1). The rat normally develops a form of
secondary hyperparathyroidism in the last several days of pregnancy,
prompted by a fall in the maternal serum-ionized calcium at the time of
rapid fetal accretion of calcium. 1,25-Dihydroxyvitamin D increases
late in gestation in rats, approximately 1 week after the rise in
intestinal calcium absorption. This indicates that mechanisms
independent of 1,25-dihydroxyvitamin D may contribute to the increased
intestinal calcium absorption in rats.
|
III. Fetal-Placental Physiology and Pathophysiology |
|---|
|
TopAbstractI. IntroductionII. Maternal Physiology and...III. Fetal-Placental Physiology...IV. Maternal Physiology and...V. Neonatal Physiology and...VI. Discussion and ConclusionsReferences |
|---|
Human handling of placental calcium transport must be largely inferred from data that have been obtained from studies in sheep, pigs, rats, and mice. Therefore, it must be emphasized that mice and rats have hemochorial placentas that are structurally very similar to those of humans (264, 265, 266, 267). In contrast, the epitheliochorial placentas of sheep and pigs differ significantly in structure from the human hemochorial placenta, and may, therefore, be functionally different as well (266).
B. Mineral ions and calcitropic hormones
1. Calcium. In humans, rodents, sheep, cattle, monkeys, and
other mammals, the fetal blood calcium (total and ionized) is
maintained at a higher level than in the maternal circulation
(268, 269, 270, 271, 272, 273, 274, 275). This elevation is mainly due to an increase in the ionized
calcium level (274). Ionized calcium is approximately 80% of the total
calcium in fetal rodents (276); only a small fraction is bound to
albumin.
In fetal rats, there is a progressive rise in total and ionized calcium over the last week of gestation, corresponding to the time of a progressive decline in fetal pH (277, 278, 279). Data are lacking on precisely how early in gestation the fetal blood calcium begins to exceed the maternal. In sheep, fetal hypercalcemia has been detected as early as the 35th day of gestation (280, 281). In humans, fetal hypercalcemia was documented at 15–20 weeks of gestation (by fetoscopy) (282) and at delivery of preterm singleton and twin pregnancies (mean gestational age 33 weeks) (283).
Two physiological models could explain fetal hypercalcemia: either the fetus maintains a fixed positive gradient of calcium with respect to the maternal level, or the fetus maintains a high, fixed level of calcium. Evidence from rat and mouse models indicates that the fetus sets its blood calcium at a higher level independently of the maternal calcium level. For example, in rats, the fetal blood calcium is unchanged in the presence of severe maternal hypocalcemia due to a calcium-restricted diet (284), vitamin D deficiency (29, 84, 285), or thyroparathyroidectomy (46, 134). The calcium gradient from mother to fetus is increased in these fetuses because the maternal blood calcium is lower. When both the pregnant rat and its fetus are thyroparathyroidectomized, the fetus still maintains a higher blood calcium level than the mother (286, 287). Also, in genetically engineered mice, maternal hypercalcemia due to heterozygous ablation of the calcium-sensing receptor (CaSR) gene does not affect the blood calcium level set by normal fetuses (288). Similarly, heterozygous calcium-sensing receptor knockout fetuses establish a constant, abnormally high fetal blood calcium level, regardless of whether the mother is heterozygous (and therefore hypercalcemic) or normal (288). The apparent "calcium gradient" is lower in offspring of these heterozygous mice, due to maternal hypercalcemia. Finally, acute alterations in the maternal blood calcium of rodents and primates (by calcium, 1,25-dihydroxyvitamin D, calcitonin, PTH, or EDTA infusions) are not reflected by much perturbation in the fetal blood calcium (232, 289, 290, 291, 292).
Others have reported a fall in the fetal blood calcium after maternal parathyroidectomy in rats (52, 53, 54). The fetal blood calcium was normal between the 12th and 17th day of gestation, but fell during the period of rapid fetal skeletal calcium accretion. Therefore, these data indicate that the ability of the fetal rat to set its blood calcium may break down during the time of rapid accretion of calcium by the skeleton, if the mother has been parathyroidectomized.
In summary, from early pregnancy, mammalian fetuses have higher levels of blood calcium than their mothers, mainly due to an increase in the ionized calcium level. The fetus does not establish a particular calcium gradient with respect to the maternal blood calcium; instead, it establishes a particular fetal blood calcium level, irrespective of the ambient maternal blood calcium level. This ability persists in the presence of significant maternal hypocalcemia of various causes, but may be impaired during the time of rapid accretion of calcium by the skeleton. The physiological importance of fetal hypercalcemia is not known.
2. Phosphate. Fetal phosphorus levels are higher than maternal in rats (279) and humans (32, 270, 273). This suggests that phosphate may be actively transported across the placenta, but the regulators of this active transport are unknown (293). PTHrP and PTH do not stimulate placental transport of phosphate in sheep (294); vitamin D may have a role (295).
3. PTH. Fetal parathyroid glands of rats and sheep contain PTH mRNA (114, 116), and PTH immunoreactivity is present in human fetal parathyroid glands as early as 10 weeks of gestation (296). These findings indicate that fetal parathyroid glands are capable of synthesizing PTH early in gestation. Furthermore, PTH detected in the fetal blood likely derives from fetal sources alone. Intact PTH does not cross the placenta of nonhuman primates, sheep, and rodents (230, 231, 232) and probably does not cross the human placenta.
The following evidence indicates that fetal parathyroid glands appear to contribute to calcium homeostasis, by secretion of PTH or PTHrP. Fetal thyroparathyroidectomy in sheep and fetal decapitation in rats caused hypocalcemia (52, 297, 298), and mice lacking the PTH/PTHrP receptor gene are hypocalcemic in utero (299). PTH can be regulated by the ambient fetal blood calcium, since EDTA-induced fetal hypocalcemia has been found to induce a rise in fetal PTH levels in rats (300), cattle (275), and rhesus monkeys (301), although another study in rhesus monkeys found no fetal PTH response (271). Removal of a maternal parathyroid adenoma was followed by a rise in amniotic fluid PTH levels and a decline in the amniotic fluid calcium level during a human pregnancy (302). Since maternal PTH cannot cross the placenta, the findings in this case have been interpreted to indicate that fetal PTH secretion can be influenced by the maternal blood calcium (302).
In fetal humans and other animals, immunoreactive PTH blood levels have been found to be undetectable or very low (i.e., <0.5 pmol/liter) with respect to maternal PTH level near the end of gestation (17, 32, 35, 38, 43, 110, 124, 268, 275, 303, 304, 305, 306, 307, 308, 309, 310). Little information is available on PTH levels earlier in gestation. One study in fetal rats found that the PTH level declined in the last several days of gestation as the serum ionized calcium rose (277), while two cross-sectional studies in preterm humans found that the fetal PTH level was not lower than the maternal PTH level (283, 308).
In summary, the available evidence suggests that the fetal parathyroids are capable of synthesizing PTH. Since blood levels of PTH have been typically found to be low in late gestation at a time when the fetal blood calcium is high, other factors must determine the fetal blood calcium level. The precise role of PTH in normal fetal calcium homeostasis will be clarified by ablating the PTH gene in mice.
4. 1,25-Dihydroxyvitamin D. Although maternal vitamin D deficiency reduces fertility and litter size in the rat (84, 85), evidence from several animal models indicates that 1,25-dihydroxyvitamin D is not necessary for normal fetal calcium and bone metabolism. In pregnant rats, sheep, and pigs that were hypocalcemic due to severe vitamin D deficiency, the fetuses maintained completely normal blood calcium and phosphate levels and had fully mineralized skeletons at term, as determined by total weight, ash weight, and calcium content of femurs (29, 84, 284, 285, 311). Each of these studies is limited by the possibility that low levels of vitamin D might have reached the fetus.
Further evidence that 1,25-dihydroxyvitamin D is not needed for normal
fetal calcium and bone homeostasis comes from the
1-hydroxylase-deficient Hannover pig model, in which the fetuses of
homozygous 1,25-dihydroxyvitamin D-deficient sows also maintained
completely normal blood calcium and phosphate levels and fully
mineralized their skeletons (79). Nephrectomy of fetal rats did not
affect fetal blood calcium or phosphate levels when measured 48 h
later, even though fetal 1,25-dihydroxyvitamin D levels fell (286).
Also in fetuses of vitamin D-deficient rats, normal levels of
calbindin28K-D and calbindin9K-D were found in
the placenta, intestine, and other tissues (132, 311, 312). In
addition, fetal mice that lack the gene encoding the receptor for
1,25-dihydroxyvitamin D are born with normal skeletons (313, 314).
Some data from humans lend support to the observation that 1,25-dihydroxyvitamin D is not needed for normal fetal calcium and bone metabolism. At term, the cord blood calcium and skeletal mineralization is completely normal in the offspring of vitamin D-deficient mothers (315, 316, 317). It is only in the first or second week after birth that hypocalcemia develops; skeletal demineralization and other rachitic changes are typically not detectable until 1 or 2 months of age (see Section V.E, below).
These observations of a minimal effect of vitamin D deficiency on fetal
calcium and skeletal metabolism do not mean that 1,25-dihydroxyvitamin
D is inactive or has no role in fetal life. In rats, the receptor for
1,25-dihydroxyvitamin D appears on day 13 of gestation in the
mesenchyme that will subsequently condense to form the skeletal
tissues, and by day 17 of gestation it is expressed in proliferating
and hypertrophic chondrocytes, and osteoblasts of limb buds and the
vertebral column (318). The widespread expression of the vitamin D
receptor early in fetal skeletal development suggests an important role
for its ligand in fetal bone development, but evidence for this
postulated role has not yet been found. Further studies manipulated the
1,25-dihydroxyvitamin D level in fetal animals to test the role of this
hormone. Infusion of antibody to 1,25-dihydroxyvitamin D decreased the
ovine fetal blood calcium level (66). 1,25-Dihydroxyvitamin D given to
pregnant guinea pigs increased the fetal calcium and phosphate levels
(319). Bilateral nephrectomy in fetal sheep resulted in reduced ionized
and total calcium and increased phosphate and PTH levels; these changes
could be reversed by administration of 1,25-dihydroxyvitamin D to the
fetus (295). Since these changes could be attributable to uremia and
not loss of the renal 1-hydroxylase enzyme, additional fetuses
underwent bilateral ureteral sectioning alone. This surgical procedure
allowed urine to drain into the fetal peritoneal cavity while retaining
functional kidneys in situ. In these fetuses, ureteral
sectioning had no effect on fetal calcium or calcitropic hormone
levels. Thus, at least in the absence of normal renal function,
1,25-dihydroxyvitamin D may have a substantial influence on fetal
mineral ion homeostasis.
1,25-Dihydroxyvitamin D does not readily cross the placenta in rats
(320); consequently, circulating levels of 1,25-dihydroxyvitamin D in
the fetus are largely derived from fetal sources. The fetal kidneys and
placenta possess the 1-hydroxylase enzyme and convert
25-hydroxyvitamin D to the active form (1, 25-dihydroxyvitamin D) (75, 76). The contribution of the fetal kidneys must be significant, since
fetal nephrectomy reduced the fetal 1,25-dihydroxyvitamin D levels in
sheep and rats (66, 286). Fetal blood levels of 1,25-dihydroxyvitamin D
are typically lower than maternal levels in humans (37, 56, 304, 321),
but umbilical artery levels of 1,25-dihydroxyvitamin D are slightly
higher than umbilical venous levels, confirming the contribution of the
fetal kidneys (37). 25-Hydroxyvitamin D levels have been found to be
roughly equal to maternal levels (37, 56); this is not surprising since
25-hydroxyvitamin D readily crosses the placenta in rats (322). Levels
of 24,25-dihydroxyvitamin D correlate with, but are typically lower
than, maternal levels at term in humans (283, 307, 321).
In summary, evidence from animal models indicates that deficiency of
1,25-dihydroxyvitamin D impairs neither fetal skeletal formation and
calcification nor the ability of the fetus to maintain a normal blood
calcium. Although these data suggest a limited role for
1,25-dihydroxyvitamin D in the fetus, fetal production of
1,25-dihydroxyvitamin D and the vitamin D receptor mandate a continued
search for fetal roles for 1,25-dihydroxyvitamin D. 25-Hydroxyvitamin D
readily crosses the placenta and can be 1-hydroxylated by the fetal
kidneys. However, 1,25-dihydroxyvitamin D does not cross the placenta,
and fetal blood levels of 1,25-dihydroxyvitamin D are low.
5. Calcitonin. Immunoreactive calcitonin can be detected in human fetal thyroid glands from as early as the 15th week of gestation (323), and fetal calcitonin levels are maintained at higher levels than maternal (35, 37, 86, 88, 89, 269, 273, 304, 324). Maternal calcitonin cannot cross the placenta (325). The increased fetal levels of calcitonin are thought to reflect increased synthesis, but the metabolism and clearance of calcitonin have not been studied in fetal animals.
Several acute experimental perturbations suggest a role for calcitonin in fetal calcium homeostasis. Infusion of calcitonin antiserum to fetal rats at day 21.5 of gestation slightly increased the fetal blood calcium 1 h later (326), while fetal injection of calcitonin caused hypocalcemia and hypophosphatemia (327). However, fetal thyroidectomy with subsequent T4 replacement did not affect the fetal blood calcium in sheep, indicating that fetal thyroidal C cells alone may not affect the regulation of the blood calcium level (298). Therefore, the precise role of calcitonin in fetal calcium homeostasis and skeletal metabolism has not yet been established.
6. PTHrP. Studies of PTH bioactivity in human umbilical cord blood (as determined by an in vitro cytochemical bioassay) found high PTH-like bioactivity, while immunoreactive PTH was simultaneously found to be undetectable or low (38, 303, 328). Subsequently, it has been recognized that human cord blood PTHrP levels are significantly higher than the simultaneous maternal levels at term (43, 310). When both PTH 1–84 and PTHrP 1–86 were simultaneously measured by two-site immunoradiometric assays [and expressed in equivalent units (picomoles/liter)], human cord blood PTHrP levels were 2–4 pmol/liter, up to 15-fold higher than the levels of PTH (0.2–0.5 pmol/liter) (43, 110, 124). It has yet to be confirmed that PTHrP accounts for the high PTH-like bioactivity in human cord blood; however, studies in fetal pigs (329) and sheep (116, 330) found that the levels of PTHrP and PTH-like bioactivity were tightly correlated in late gestation and the neonatal period.
As noted earlier (Section II.B.6), PTHrP may be a prohormone that is processed into separate circulating fragments, each of which may have different functional roles and receptors (100). Although the structures of these fragments have been deduced from studies of tumor cell lines transfected with the PTHrP gene, it has yet to be determined which of these fragments normally circulate in fetal life. Full-length PTHrP has twice the molecular weight of PTH; since PTH cannot cross the placenta, PTHrP probably does not either. PTHrP 1–86 did not cross the placentas of sheep and goats (113); however, the possibility that smaller, biologically active fragments of PTHrP might cross the placenta has not been evaluated.
PTHrP is produced in many sites throughout the developing embryo and fetus, including the fetal parathyroid glands (116, 331), skeletal growth plate (332, 333), trophoblast cells of the placenta (114, 331), amnion (106, 334), chorion (334), umbilical cord (115), and many other organs. All of these sites may contribute to the circulating level of PTHrP in the fetus and may thereby be relevant to fetal calcium and bone metabolism. Since venous umbilical PTHrP levels were higher than umbilical arterial levels in pigs, the placenta may be an important source of systemically circulating PTHrP in the fetus (329). Due to local production of PTHrP by the umbilical cord (115), the level of PTHrP in cord blood might not accurately reflect the systemic level of PTHrP, but this has not been tested.
PTHrP has multiple possible roles during embryonic and fetal development (335). PTHrP gene-ablated mice have abnormalities of chondrocyte differentiation (336) and aberrant breast development (337). PTHrP may also be an important regulator of the fetal blood calcium. PTHrP levels correlate with the fetal ionized calcium levels in pigs (329). In gene-tically engineered mice, homozygous ablation of the PTHrP gene results in a fetal blood calcium no higher than that of the mother (299). In sheep, fetal parathyroidectomy causes hypocalcemia that can be reversed by PTH or PTHrP infusion (297, 298, 338). Since PTH normally circulates at low or undetectable levels in the fetus near term (Section III.B.3, above), it is possible that the hypocalcemic effect of fetal parathyroidectomy is at least partly due to the loss of PTHrP produced by the parathyroids. In the next section (III.C), the unique role of PTHrP in stimulating placental calcium transport will be discussed.
In summary, PTHrP is produced by diverse fetal tissues and circulates in fetal blood at levels higher than adult levels. PTHrP appears to regulate the fetal blood calcium as well as fetal-placental calcium transport.
C. Fetal-placental calcium transport
Calcium is actively transported across the placenta (339, 340).
The site of active transport is likely at the fetus-facing basement
membrane of the syncytiotrophoblast cells in the human and at the
trophoblast cells and the basal surface of the endoderm of the
intraplacental yolk sac in rodents (341, 342). The active transport of
calcium may be mediated by a Ca2+-ATPase present at these
same sites (339, 342). This enzyme’s activity can be inhibited by
dinitrophenol, ouabain, quercetin, and antibody to the human
erythrocyte plasma membrane calcium pump (339, 342). Calcium-binding
proteins in the placenta and yolk sac are also thought to be involved
in active placental calcium transport. The placental
calbindin9K-D mRNA and protein levels increase over the
last week of gestation in rats (129, 343) and mice (130, 133) and are
unaffected by maternal vitamin D deficiency (132, 311). Transplacental
transport of calcium is generally found to be a one-way process,
i.e., fetal-to-maternal flow of calcium is typically less
than 1% of the forward (maternal-to-fetal) flow (344, 345). In rhesus
monkeys, backflux was reported to be 80% of the forward flow (345); it
is not certain whether this represented a true species difference or
methodological differences. It has not been determined when active
transport of calcium begins in gestation, due to technical difficulties
involved in studying placental physiology early in gestation. However,
active transport of calcium must be underway by the third trimester in
humans, which is the time of rapid skeletal mineralization and peak
fetal calcium requirement.
1. Maternal hormones. Maternal hormones might influence the fetal-placental calcium transport process by raising or lowering the ambient maternal calcium level, and by direct effects on the placenta. However, several lines of evidence from animal experiments indicate that fetal-placental calcium transport and net maternal-fetal calcium transfer are maintained relatively independently of maternal hypocalcemia or hormone deficiencies. In pregnant sheep, maternal hypocalcemia due to parathyroidectomy or dietary calcium restriction did not affect the rate of fetal-placental calcium transfer as directly assessed in placental perfusion experiments (297, 346); in addition, the fetal blood calcium, phosphate, PTH, and 1,25-dihydroxyvitamin D levels were all unchanged by maternal hypocalcemia (284, 347). The finding of a "normal" rate of calcium transfer across the placenta indicates that the fetal-placental unit must be working harder to extract the normal amount of calcium from a reduced amount of maternal calcium presented to the placenta. Indeed, the following observation from intact fetal rats confirmed that the rate of placental calcium transfer is up-regulated in response to
