Placental-specific Igf2 knockout mice exhibit hypocalcemia and adaptive changes in placental calcium transport

Abstract

Evidence is emerging that the ability of the placenta to supply nutrients to the developing fetus adapts according to fetal demand. To examine this adaptation further, we tested the hypothesis that placental maternofetal transport of calcium adapts according to fetal calcium requirements. We used a mouse model of fetal growth restriction, the placental-specific Igf2 knockout (P0) mouse, shown previously to transiently adapt placental System-A amino acid transporter activity relative to fetal growth. Fetal and placental weights in P0 mice were reduced when compared with WT at both embryonic day 17 (E17) and E19. Ionized calcium concentration [Ca²⁺] was significantly lower in P0 fetal blood compared with both WT and maternal blood at E17 and E19, reflecting a reversal of the fetomaternal [Ca²⁺] gradient. Fetal calcium content was reduced in P0 mice at E17 but not at E19. Unidirectional maternofetal calcium clearance (^Ca K _mf) was not different between WT and P0 at E17 but increased in P0 at E19. Expression of the intracellular calcium-binding protein calbindin-D_9K, previously shown to be rate-limiting for calcium transport, was increased in P0 relative to WT placentas between E17 and E19. These data show an increased placental transport of calcium from E17 to E19 in P0 compared to WT. We suggest that this is an adaptation in response to the reduced fetal calcium accumulation earlier in gestation and speculate that the ability of the placenta to adapt its supply capacity according to fetal demand may stretch across other essential nutrients.

Fetal growth restriction (FGR), the failure of a fetus to achieve its genetic growth potential, significantly increases the risk of infant mortality and morbidity (1): 5 to 10% of all pregnancies complicated by FGR result in stillbirth (2, 3) and around 5% of the survivors of FGR go on to develop cerebral palsy (3), with 8% suffering some form of neurological impairment (3). Furthermore, there is a now well-defined relationship between low birth weight per se and increased risk of diseases in adulthood, such as hypertension, type-2 diabetes (4), and osteoporosis (5).

The major pathway of nutrient transfer to the fetus is via the placenta (reviewed in ref. 6). Therefore, the net flux of nutrients across the placenta over an entire gestation must be lower in a baby of low birth weight when compared to a larger infant. Although a reduced net flux might result from a number of factors, such as poor maternal nutrition, and therefore reduced maternal plasma concentration of specific solutes, the major cause of FGR in the developed world is placental insufficiency (7). It is now clear that all of the factors that determine the capacity of the placenta to transfer nutrients, including blood flow, exchange of barrier structure, and the activity of specific nutrient transporters, may be abnormal in FGR (8 –16). Recent data suggest (reviewed in ref. 17) that of these placental changes, some are likely to be causative and others to be adaptive to the growth restriction.

The concept that elements of placental function adapt to maintain the supply of nutrients appropriate to the genetic growth potential of the fetus, in the face of other placental maldevelopments, is unique and largely unexplored. This idea arises from the synthesis of separate observations in human and mouse, focusing on the activity of the System A amino acid transporter. The activity of this transporter (per milligram of membrane protein) in vesicles isolated from the microvillous (maternal facing) plasma membrane (MVM) of the syncytiotrophoblast (transporting epithelium) of the human placenta, was found to be inversely related to birth weight and size of the baby across the range of normal birth weights and size (18). In contrast, in MVM from placentas of FGR babies, activity was significantly lower than that in MVM from placentas of normally grown babies (13). An interpretation of these data are that there is an up-regulation of System A activity per unit membrane in the placentas of small normal babies, increasing placental transport efficiency and maintaining fetal growth. Failure of such an adaptive up-regulation of System A activity in FGR could be a cause of the diminished fetal growth (13, 18).

Further studies have examined the activity of System A amino acid transporter in mice, where the placental-specific transcript (P0) of the insulin-like growth factor-2 (Igf2) gene has been deleted. Although placentas from P0 knockout mice are smaller than wild type (WT) mice from about embryonic day 12 (E12), fetuses are not significantly smaller until around E18 (term is E20). Maternofetal transfer of ¹⁴C-methylaminoisobutyric acid (a specific, nonmetabolisable amino acid substrate of System A), measured in vivo, was significantly higher, per gram of placenta, in the P0 knockouts when compared with WT, at E16 but not at E19 (9). Similarly, mRNA expression of the slc38a4 isoform of the System A transporter gene was higher in the placentas of P0 knockout mice than WT at E16 but not at E19 (8). These data are consistent with human studies, suggesting that when placental growth is restricted early in gestation, System A is up-regulated to meet fetal nutrient demand, but this adaptation is not maintained until term, contributing to the ensuing FGR. In a different experimental paradigm, Coan et al. (19) used normal mice to investigate whether the smallest placentas in a litter are more efficient in supplying fetal nutrients than larger placentas. They found morphological adaptations and increased System A amino acid transporter activity and mRNA expression in the smallest placentas, which were postulated to increase placental transport capacity and maintain normal fetal growth. Such functional adaptations in these situations may well be modulated by the prevailing endocrine environment of the fetus and the placenta, driving the resultant increase in placental efficiency (20).

In human FGR, the activity of a number of different placental solute transporters has now been investigated and compared with that in normal pregnancy (17). Various changes in transporter activities have been observed in FGR: some lowered (e.g., System A, as described above), some not changing (e.g., glucose transporter), and interestingly one transporter, Ca²⁺ATPase, showing increased activity in the basal (fetal facing) plasma membrane of the syncytiotrophoblast. This observation that placental calcium transporter activity is up-regulated in FGR led to our speculation that this is an adaptive response, a concept we have investigated in this study using the P0 mouse as a model of FGR.

Calcium transport across the placenta is an active process, with fetal plasma ionized calcium concentration ([Ca²⁺]) being significantly higher than maternal [Ca²⁺] at term (6). Transplacental calcium transport is believed to have three components (6): influx of calcium from maternal blood into the trophoblast cytosol via the transient receptor potential vanilloid 6 (TRPV6) channel, translocation across the cytosol on a binding protein (which in rodents is thought to be calbindin-D_9K), and active efflux out of the trophoblast into fetal-side placental extracellular fluid via plasma membrane Ca²⁺ATPase (PMCA). In rodents, however, two potential routes exist for maternofetal calcium transfer: (i) calcium transport across the labyrinth trophoblast, considered to be the predominant site of maternofetal exchange, and (ii) the intraplacental yolk sac (IPYS). The IPYS is an exclusive structure formed in rodent placenta by the invagination of the primitive yolk sac into the chorioallantoic placenta (21). The high density of calcium transporters, calcitropic hormones, and receptors in the IPYS (21, 22) strongly implies that this structure plays an important role in maternofetal calcium exchange and the provision of calcium to the developing fetus (21).

We hypothesized that the increase in Ca²⁺ATPase activity in FGR was a direct example of a placental adaptation, which could serve to promote fetal calcium acquisition despite the small size and malformation of the FGR placenta. This hypothesis has been tested in the present study by exploiting the well-characterized FGR phenotype of the P0 knockout mouse (hereafter referred to as P0), a mouse model that allowed all aspects of placental calcium transfer as well as fetal calcium accretion to be measured, and which exhibits evidence of adaptation with respect to another placental transport system, namely System A.

To investigate our hypothesis, we measured [Ca²⁺] in maternal and fetal plasma, total fetal calcium accretion (as an estimate of net placental calcium flux), and the unidirectional maternofetal clearance of ⁴⁵CaCl₂. In this study, all measurements were carried out at both E17 and E19, the period over which maternofetal calcium transport increases exponentially and skeletal mineralization occurs, and comparisons were drawn between P0 fetuses and WT siblings in the same litter.

Results and Discussion

Placental and fetal weights and fetal:placental weight ratios for P0 and WT fetuses are shown in Table 1. At both E17 and E19, fetal and placental weights were significantly reduced in P0 as compared to their WT siblings (P < 0.001). Fetal:placental weight ratios were significantly increased in P0 mice at both gestational ages (P < 0.001). This finding confirms previous observations that P0 mice demonstrate growth restriction toward term following a significant reduction in placental weight that, according to previous reports, occurs earlier, at around E14 (8, 9). This chronology indicates placental insufficiency as the major cause of FGR in these mice.

Table 1.

Fetal and placental weights and fetal:placental weight (F:P) ratio in WT and Igf2 P0 promoter (P0) knockout mice at E17 and E19

	E17		E19
	WT	P0	WT	P0
Fetal weight (g)	0.60 ± 0.01	0.54 ± 0.01*	1.20 ± 0.02	0.93 ± 0.02*
Placental weight (g)	0.095 ± 0.001	0.066 ± 0.001*	0.094 ± 0.003	0.062 ± 0.002*
F:P ratio	6.4 ± 0.1	8.2 ± 0.2*	13.0 ± 0.4	15.2 ± 0.5*

Data are expressed as mean ± SEM of the average placental and fetal weights of WT and P0 fetuses from within an individual litter, where n = 15 and 16 litters at E17 and E19, respectively.

*P < 0.001 vs. WT at same gestational age (paired t test).

[Ca²⁺] in fetal blood of WT and P0 fetuses at E17 and E19 is shown in Fig. 1. At E17, fetal [Ca²⁺] in WT fetuses was not significantly different from that in maternal blood (≈1.1 mmol/L) indicating that a fetomaternal [Ca²⁺] gradient was not yet established. However, at E19 fetal [Ca²⁺] was significantly higher than maternal [Ca²⁺]. These data demonstrate that the fetomaternal calcium gradient in these mice is established between E17 and E19, and further suggest that active transplacental calcium-transport mechanisms during this period begin to contribute significantly to maternofetal calcium flux, consistent with other studies in mice and rats (23 –26). In contrast, fetal blood [Ca²⁺] in P0 was significantly lower than WT and maternal blood at both E17 and E19 (P < 0.05). This failure of P0 fetuses to achieve fetal hypercalcemia relative to the mother close to term could reflect an increased removal of fetal plasma calcium to the fetal skeleton or, alternatively, reduced net placental calcium transport.

Fig. 1.

Fetal [Ca²⁺] in WT and P0 fetuses was measured in whole fetal blood at E17 (A) and E19 (B). Dotted line denotes mean [Ca²⁺] in maternal blood (n = 10 and n = 12 at E17 and E19, respectively). Each dot represents the mean of either WT or P0 within a single litter. n = 6 litters at E17, n = 7 litters at E19. *, P < 0.05, Wilcoxon signed-rank test.

To explore these possibilities, we measured fetal calcium content [as an estimate of net transplacental calcium flux (^Ca J _net)] (27) and unidirectional maternofetal clearance of ⁴⁵Ca (^45Ca K _mf) across the intact placenta. Fetal calcium content (mmol/g ash weight) of WT and P0 fetuses at E17 and E19 is shown in Fig. 2. For ease of comparison, data are expressed as a ratio of P0/WT. At E17, P0 fetal calcium content was lower compared to WT siblings in all seven litters examined (P < 0.01). At E19, this difference was no longer apparent with P0 and WT demonstrating comparable fetal calcium content. In taking fetal calcium content as an estimate of ^Ca J _net (27), these data suggest that ^Ca J _net in P0 fetuses had been lower up to E17, as compared with WT fetuses, but by E19 this had been normalized in P0 fetuses, implying that ^Ca J _net in P0 fetuses had been stimulated over the period of E17 to E19. ^Ca J _net will reflect the difference between unidirectional maternofetal (^Ca J _mf) and fetomaternal (^Ca J _fm) calcium fluxes to and from the fetus (27), which are often measured as respective unidirectional clearances (23, 27). We have previously demonstrated that ^Ca J _mf approximates to ^Ca J _net at E18 in WT mouse fetuses and that ^Ca K _mf is more than 70-times greater than the unidirectional fetomaternal clearance of calcium (^Ca K _fm) (23). We therefore examined whether the increase in calcium accretion in P0 fetuses between E17 and E19 was attributable to a change in ^Ca J _mf, as measured by ^Ca K _mf (23, 27), using ⁴⁵CaCl₂ as tracer.

Fig. 2.

Fetal calcium content in WT and P0 fetuses at E17 and E19. Each symbol corresponds to fetal calcium content (measured as mmol per gram of ash weight) expressed as a ratio of P0/WT within a single litter. Solid line denotes median values for all of the litters shown at each gestational age. Dotted line represents ratio of 1 (no difference between P0 and WT). Absolute values expressed as mmol/g ash weight (median, interquartile range) are as follows: E17 WT (1.51, 1.35–2.14), E17 P0 (1.39, 1.18–1.75), E19 WT (4.02, 3.90–4.44), and E19 P0 (4.01, 3.26–4.30). **, P < 0.01, Wilcoxon signed-rank test.

Fig. 3 shows ^45Ca K _mf measured across the intact placenta and expressed as a P0/WT ratio. At E17, there was no difference in ^45Ca K _mf between genotypes. However, at E19 the P0/WT ratio of ^45Ca K _mf was significantly above 1, indicative of an increased ^45Ca K _mf in P0 fetuses as compared with their WT siblings. This increased ^45Ca K _mf is consistent with an adaptive response to the reduced fetal calcium content at E17, acting to restore P0 fetal calcium content, per gram of fetal ash, to a comparable level to WT at E19.

Fig. 3.

Unidirectional maternofetal clearance of calcium (^Ca K _mf) across placentas of WT and P0 fetuses at E17 and E19. Each symbol corresponds to ^Ca K _mf (μL/min per gram placenta) expressed as a ratio of P0/WT within a single litter. Solid line denotes median values for all of the litters shown at each gestational age. Dotted line represents ratio of 1 (no difference between P0 and WT). Absolute values expressed as μL/min per gram placenta (median, interquartile range) are as follows: E17 WT (103.8, 87.2–156.7), E17 P0 (102.0, 82.1–161.0), E19 WT (79.1, 20.4–169.7), and E19 P0 (105.1, 47.7–239.8). **,P < 0.01, Wilcoxon signed-rank test.

Finally, we examined whether these changes in placental calcium flux are associated with changes in the expression of proteins involved in transplacental calcium transfer. Fig. 4 shows the protein expression of TRPV6, calbindin-D_9K, and PMCA1 by Western blotting. All proteins were observed at the predicted sizes, consistent with previous observations (23). As would be predicted, TRPV6 was highly expressed in the first plasma membrane barrier to restrict solute transfer in mouse placenta: namely, the maternal-facing plasma membrane of trophoblast layer II included here as a positive control (lane mAM in Fig. 4A). Evidence suggests that this plasma membrane shares functional homology with MVM of the syncytiotrophoblast in human placenta (28). β-actin, used as a loading control, demonstrated equal loading between groups. Placental TRPV6 expression, recently shown to be important in maternofetal calcium transport (29), was not different between P0 and WT at either E17 or E19. In contrast, calbindin-D_9K expression was significantly lower in P0 placentas compared to WT at E17 (P < 0.01), with this trend observed in all eight litters. However, by E19 calbindin-D_9K expression was comparable between the two genotypes. This mirroring of the trend observed in the fetal calcium content of P0 fetuses and the placental expression of calbindin-D_9K over days E17 to E19 lends further support for the concept that calbindin-D_9K expression is rate-limiting for fetal calcium acquisition in rodents (23, 24, 30, 31).

Fig. 4.

Protein expression of TRPV6, calbindin-D_9K and PMCA1 in placentas of WT and P0 fetuses at E17 and E19. Representative blots show expression within four paired litters. n = 8 litters at E17 and n = 6 litters at E19, except for calbindin-D_9K at E19, where n = 8 litters. β-actin was used as a marker of equal loading between samples and a single blot that has been stripped and reprobed for β-actin is shown. Arrows denote size of placental immunoreactive proteins. Protein loading was as follows: TRPV6, 20 μg placental membrane; calbindin-D_9K, 50 μg postnuclear supernatant at E17, 20 μg at E19; PMCA1, 40 μg placental membrane. mAM and RB denote mouse placental apical membrane (derived from maternal facing-plasma membrane of syncytiotrophoblast layer II) and rat brain, respectively, included as positive controls. Graphs show densitometry values for TRPV6 (A), calbindin-D_9K (B), and PMCA1 (C) at both E17 and E19. Each symbol corresponds to densitometry values expressed as a ratio of P0/WT within a single litter. Solid line denotes median values for all of the litters shown at each gestational age. Dotted line represents ratio of 1 (no difference between P0 and WT). **, P < 0.01, Wilcoxon signed-rank test.

PMCA1 expression was not different between P0 and WT at either E17 or E19. However, in human placenta PMCA protein expression and activity are not always correlated; in FGR increased activity of the PMCA transporters (measured in syncytiotrophoblast basal-plasma membranes) was associated with a reduced protein expression of PMCA (10). We are not able to address this issue further in the present study, as there is currently no method to isolate the fetal-facing plasma membrane from trophoblast layer III of mouse placenta, likely to be analogous to the basal syncytiotrophoblast plasma membrane in human placenta, based on PMCA localization to this plasma membrane in rodent placenta (32).

Our data show that total calcium accretion by P0 fetuses is lower than that of corresponding WT siblings at E17. The reasons for this diminished accretion are not clear at present. In considering calcium fluxes, it would have been predicted that ^Ca J _mf (as reflected by ^45Ca K _mf) would have been lower across the placentas of P0 fetuses at E17. However, at E17 we found that ^45Ca K _mf was not different between P0 and WT fetuses, although calbindin-D_9K expression was significantly reduced in the placentas of P0 fetuses, entirely consistent with their reduced fetal calcium content. This discordance between placental calbindin-D_9K expression and maternofetal calcium flux was surprising in the light of other evidence showing a good correspondence between these variables under different situations (23, 24, 30). This finding could relate to the gestational timing of our observations, as there is a highly exponential gestational increase in ^Ca J _mf between E17 and E19 in rodents (23, 30). Although this gestational rise in ^Ca J _mf shows a strong stoichiometric relationship with placental calbindin-D_9K expression (30), onset of the dynamic changes in ^Ca J _mf may have already commenced in P0, preceding the modulation of calbindin-D_9K expression. It would be interesting to investigate ^Ca J _mf in P0 fetuses earlier in gestation, but this may not provide further elucidation, bearing in mind that fetal skeletal mineralization occurs over the last few days of gestation in rodents and placental calbindin-D_9K expression is comparatively low before E15 (30, 33–36). It is also possible that activity of PMCA was increased in the placentas of P0 fetuses at E17, but as alluded to above, direct investigation of this issue is not easily accomplished. Another possibility is that ^Ca J _fm, which is very difficult to measure across the intact placenta in vivo, is higher in P0 than WT fetuses during early gestation, thus reducing ^Ca J_net. However, we have previously measured ^Ca K _fm near term in the mouse, and this is less than 1.4% of the magnitude of ^45CaK_mf (23), suggesting that this flux would have to increase by at least an order of magnitude in P0 mice to cause a significant reduction in ^Ca J _net, which seems unlikely. Finally, lower calcium accretion in the P0 fetuses might reflect an altered calcium flux across the choriovitelline (or yolk sac) placenta early in gestation, before the chorioallantoic placenta forms [by E10 (37)]. This notion would accord with ^Ca J _mf being the predominant flux across the visceral yolk sac of other species and approximating to ^Ca J _net (38). It should also be commented upon that although placental calbindin-D_9K expression is reduced in P0 at E17, the cellular origin of this response is uncertain, with a relatively high abundance of both calbindin-D_9K and PMCA within the mouse IPYS, along with calcitropic hormomes and receptors, providing a potential maternofetal calcium-transfer pathway in addition to exchange across the labyrinth trophoblast (21, 22). Considering the marked induction in placental calbindin-D_9K expression in the IPYS toward term (39), one may suggest that any reduction in calbindin-D_9K expression within the IPYS of P0 placentas at E17 would have significant effects on fetal calcium accretion. Certainly, the relative contributions of the IPYS and labyrinthine trophoblasts toward this reduced calbindin-D_9K expression is worthy of further investigation.

Whatever the causes of the lower calcium accretion at E17 in the P0 fetuses, it is clear that around this time it is sensed in such a way as to lead to a corrective response by the fetoplacental unit. Thus, placental calcium transport adapts and responds with increased relative calbindin-D_9K expression and ^Ca J _mf, so that by E19, P0 fetal calcium content (mmol/g ash weight) is normal. Therefore, despite the marked growth restriction of the P0 fetuses at both E17 and E19 and a reduced placental mass, skeletal hypomineralization of the P0 fetus at E17 is corrected by E19, and this is achieved by a rise in ^Ca J _mf. Work from several species suggests that active placental calcium transport is switched on, or up-regulated, near to term, as bone mineral accretion begins (23, 25, 26, 40–43). The capacity to raise ^Ca J _mf in P0 indicates that the functional integrity of active calcium-transport mechanisms in P0 placentas is preserved. In the light of this finding, it is perhaps surprising that P0 fetuses remain hypocalcemic and fetal [Ca²⁺] does not rise alongside that seen in WT fetuses to become hypercalcemic relative to the mother. This finding presumably follows from the increased calcium uptake into the skeleton of P0 fetuses reflected in their normal calcium content by E19.

It is noteworthy that the ontogeny of the adaptive response to placental calcium transport in P0 fetuses shown here is the opposite to the trend we have previously observed for System A amino acid transport; this was higher than WT at E16 and then decreased toward term (9). These data accord with previous observations in human placenta near term showing that Ca²⁺ATPase activity is increased even when System A amino acid transporter activity is decreased in FGR-affected pregnancies (10, 13). Interestingly, in the mouse placenta there appears to be a gestational sequence to the adaptive responses improving placental efficiency, with morphological changes preceding those observed in amino acid transporter activity (19). Together, these data highlight that the molecular mechanisms involved in the placental adaptive response are complex and temporally regulated. The fetal “demand” signals to elicit these changes in placental function remain to be fully elucidated, but are likely to be multifactorial. As regards placental calcium transport, the increasingly well-described role of parathyroid hormone-related peptide in regulating placental calcium transport (23, 25, 26, 43) suggests one attractive candidate. This finding again highlights the potential role of the IPYS in the regulation of placental calcium transport, as both the PTH/PTHrP receptor and PTHrP are known to be highly expressed within the IPYS (21).

In summary, this study provides direct evidence that the capacity of the placenta to supply a particular nutrient is adaptable in relation to the fetal demand or requirement for that solute. Our data also suggest, importantly, that there are nutrient-, and perhaps gestation-specific signals operating between the fetus and placenta to enable any particular adaptation. As has been recently pointed out by others (19), a variable sequence of placental adaptations occurring at different times in gestation will have marked affects on nutrient delivery to the fetus and, ultimately, the composition of fetal tissues, resulting in programming an individual’s homeostatic mechanisms with metabolic consequences extending into adulthood. In this context, and of particular relevance to this study, is the association between birth weight and osteoporosis in adult life (5). Clearly, this association between reduced bone-mineral content in later life and low birth weight puts a high dependence on the placenta to provide adequate calcium provision to the developing fetus. By elucidating the gestational-specific changes in ^Ca J _mf and fetal calcium accretion in this study, it is hoped that placental adaptations in calcium transport in human FGR may begin to become better understood. Understanding the link between placental phenotypes arising from adaptations to altered fetal-growth trajectories may therefore be crucially important.

Materials and Methods

Animals.

Experiments were performed in accordance with the U.K. Animals (Scientific Procedures) Act of 1986. Placental specific Igf2 (Delta U2 P0) knockout mice, which had deletion of the U2 exon of the Igf2 gene, were generated as previously described (44) and were a kind gift from W. Reik (Babraham Institute, Cambridge, UK).

Wild-type C57/BL female mice (8–12 weeks old) and males heterozygous for the P0 deletion (6 weeks to 8 months old) were mated and the first day of gestation determined by the discovery of a copulation plug (term = 19–20 days). All animals were provided with nesting material and housed in cages maintained under a constant 12-h light/dark cycle at 21 to 23 °C, with free access to food (Beekay Rat and Mouse Diet; Bantin & Kingman) and tap water.

Genotyping.

Fetuses were genotyped using genomic DNA extracted from fetal tail tips (DNeasy, Qiagen). Igf2 P0^+/− mutants were identified using a primer pair to amplify a 740-bp fragment across the 5-kb deletion (sense, 5′-TCCTGTACCTCCTAACTACCAC -3′; antisense, 5′-GAGCCAGAAGCAAACT -3′) with a third primer amplifying a 495-bp fragment from the WT allele (5′- CAATCTGCTCCTGCCTG-3′), as described previously (9).

Unidirectional Maternofetal Clearance of ⁴⁵Ca Across the Intact Placenta (^Ca K _mf).

^Ca K _mf across the intact placenta was measured at E17 and E19 using an adaptation of the method of Flexner and Pohl, as described previously (23). Following infusion of ⁴⁵Ca, exsanguination of the dam occurred between 1- and 6-min postinfusion in accordance with previous studies (23). Fetuses were rapidly collected and assessed for total radiolabel accumulation and compared to a maternal plasma ⁴⁵Ca disappearance curve (see below).

^Ca K _mf was calculated as:

Where, N_x = total radiolabel accumulation (dpm) by the fetus (corrected for the fetal tail tip retained for genotyping) at x min after injection of radiolabel into the maternal vein, W = placental wet weight (g) and = the time integral of radioisotope concentration in maternal plasma (dpm × min)/μL from 0 to x min (taken from the maternal plasma ⁴⁵Ca disappearance curve).

Maternal Plasma ⁴⁵Ca Disappearance Curve Following Injection into the Maternal Circulation.

A maternal plasma ⁴⁵Ca disappearance curve was constructed from dams at either E17 (n = 20) or E19 (n = 37) and fitted to a one-phase exponential decay model (r ² > 0.5), as described previously (23).

Fetal Calcium Content.

Calcium content of fetal ash was measured at E17 and E19 by atomic absorption spectrophotometry (Solaar S-Series, Thermo Elemental, Unicam Ltd.), as described previously (23). Tail tips were taken for genotyping as described above.

Whole-Blood Ionized Ca²⁺ Concentration.

Terminal blood samples, taken from either mother or fetus at E17 and E19 from the carotid artery following decapitation, were analyzed for ionized Ca²⁺ concentration (Bayer 865 Analyzer, Siemens). Ionized Ca²⁺ measurements were corrected to pH 7.4. No correction was made for concentration of plasma proteins within individual samples.

Western Blotting.

Individual placentas harvested at E17 and E19 were homogenized in buffer containing 300 mM mannitol, 12 mM hepes (pH 7.6), and 1% protease inhibitor mixture (Sigma-Aldrich) for 30 s. The homogenate was retained or centrifuged at 2,500 × g for 5 min at 4 °C (Sorvall Discovery 100SE; Kendro Laboratory Products). Aliquots of this postnuclear supernatant were retained and the remaining postnuclear supernatant centrifuged at 100,000 × g for 30 min at 4 °C to obtain the membrane fraction. Both fractions were analyzed for protein concentration (Bio-Rad Protein Assay) and stored at −80 °C for further analysis.

Protein-SDS complexes were prepared with or without heat denaturation (23) and SDS/PAGE performed followed by electrotransfer to nitrocellulose membranes (GE Healthcare). The antisera used were rabbit polyclonal anti-human TRPV6 (1:200; Santa Cruz Biotechnology, Insight Biotechnology), rabbit polyclonal anti-rat calbindin-D_9K (1:1,000; SWANT), rabbit polyclonal anti-human plasma membrane calcium ATPase isoform 1 (PMCA1, 1:500; SWANT), and rabbit polyclonal anti-human β-actin (1:1,000; Abcam). Negative controls were prepared by omission of primary antibody. All gels used were 7% acrylamide with the exception of calbindin-D_9K, where a 15% gel was used because of the relatively small size of the target protein. Immunoreactive species were detected with horseradish peroxidase-conjugated secondary antibodies (1:2,000, Dako) using an enhanced chemiluminescence detection system (GE Healthcare). Immunoreactive signal density was measured by densitometry (Image J, National Institutes of Health) and all signals fell within the linear range of detection.

Statistical Analysis.

All data are presented as either mean ± SEM or as individual dot plots per litter, where n = number of litters. For most parameters, either a paired t test or Wilcoxon signed-rank test were used to test for statistical differences between groups, dependent on normal distribution of data. P < 0.05 was taken as the significance level.