Dietary calcium deficiency in laying ducks impairs eggshell quality by suppressing shell biomineralization

The eggshell of birds plays an important role in determining the physical and antimicrobial defense of the egg and regulates the exchange of metabolic gases and water (Hincke et al., 2012). Eggshell formation is a calcium-demanding process that takes most of the time (20 h) during the whole process of egg formation (25–26 h). It occurs in the uterus (shell gland), where calcite crystals firstly nucleate on specific sites (mammillary knobs) composed of organic cores and then the various layers of the eggshell assemble sequentially (Nys et al., 2004; Gautron and Nys, 2006; Rodríguez-Navarro et al., 2015).

The process of shell biomineralization involves calcium carbonate synthesis and calcite crystal deposition (Nys et al., 2004). The calcium (Ca²⁺) and bicarbonate (HCO₃⁻) that are supplied by the blood via trans-epithelial transport (Arias, et al., 1991,, 1997) are the most important prerequisites for shell biomineralization in the uterine lumen. Ionized calcium and bicarbonate concentrations in uterine fluid remain supersaturated for calcite solubility and ensure the process of shell calcification (Nys et al., 1991). Because the eggshell contains 95% calcium carbonate (weight/weight), approximately 5–6 g of calcium carbonate is deposited, in its calcitic form, in the chicken eggshell during the stage of calcification (Creger et al., 1976; Stemberger et al., 1977). The transition from amorphous particles of calcium carbonate to calcite crystals, also called calcification, is facilitated by matrix proteins in the uterus (Hernández-Hernández et al., 2008; Rodríguez-Navarro et al., 2015) and involves three different stages of nucleation and deposition of calcite crystals: initiation, fast growth and termination. Studies have indicated that the matrix proteins determine the crystal morphology (Weiner and Addadi, 1991; Lakshminarayanan et al., 2005), eggshell structure (Freeman et al., 2010) and eggshell mechanical properties (Dunn et al., 2009). More than 500 proteins have been identified in the chicken eggshell proteome (Mann et al., 2006). The occurrence of some of these matrix proteins in the uterus coincides with the concomitant presence of the egg in each uterine region (Lavelin et al., 1998,, 2000; Brionne et al., 2014), suggesting an important role for the matrix protein in the control of eggshell calcification.

Although it is known that dietary calcium represents the most important material for eggshell formation, the underlying molecular role of calcium in shell biomineralization has not been fully understood. The objective of the present study was to evaluate the changes in key proteins involved in shell formation by depletion and subsequent repletion of dietary calcium provided to laying ducks.

After 67 days of calcium depletion, egg production was reduced to 84% or 47% of control for 1.8% calcium and 0.38% calcium, respectively (Fig. 1). Similarly, egg mass, average shell thickness and relative shell weight from ducks consuming the 0.38% calcium diet were significantly decreased from those in the calcium-adequate controls (Table 1). A significant reduction in the relative shell weight was also observed in ducks consuming 1.8% or 0.38% calcium (P<0.0001). Similarly, shell breaking strength decreased with decreasing dietary calcium. These differences were all eliminated after ducks were made calcium replete during the second 67-days period on the control diet (Table 1).

After calcium depletion, both the mammillae density and palisade layer thickness were significantly reduced in ducks fed 0.38% calcium (P<0.01) (Fig. 2F,I) but not in ducks fed 1.8% calcium (Fig. 2E,H; Table 2). Similarly, reduced mammillary knob thickness and total thickness of the mammillary knob and palisade layer were also observed in both 1.8% calcium and 3.6% calcium (Table 2, Fig. 2). There were no differences in any parameters of the microstructure of the eggshell among treatments after all ducks were fed with the adequate calcium diet for 67 days.

Free calcium ions in plasma were reduced to about 87% of levels in controls (P<0.05) after ducks had been fed either of the calcium-deficient diets (1.8% or 0.38%) for 67 days (Table 3). The decrease in uterine calcium was more striking: to 68% (1.8% calcium diet) and 55% (0.38% calcium diet) of control levels. Although still significant, the decreases in calcium content of the eggshell when ducks were fed less than adequate dietary calcium were only about 5% (Table 3). There were no differences in any of these indices of calcium status after the second phase of feeding, when all ducks had received the adequate calcium diet for 67 days. The phosphorus level in plasma, uterus and eggshell were not affected during the period of calcium depletion and repletion.

The relative abundance of both SPP1 and carbonic anhydrase 2 (CA2) mRNA transcripts in the uterus decreased after feeding calcium-deficient diets (1.8% or 0.38% calcium). Transcripts of calbindin 1 (CALB1), ovocalyxin-32 (OCX-32) and ovocalyxin-116 (OC-116) were reduced in ducks fed the 0.38% calcium but not the 1.8% calcium diet (Fig. 3). By contrast, clusterin (CLU) mRNA increased in ducks fed 1.8% (P<0.05) or 0.38% (P<0.0001) dietary calcium, compared with the controls (Fig. 3). The relative expression of ovalbumin (OVA) was unchanged. After calcium repletion for the subsequent 67 days, there were no significant differences in relative abundance of any of these transcripts, although some, for example, CALB1, OC-116 and CLU, tended to overshoot values for the controls (Fig. 3).

Plasma estradiol concentrations were significantly decreased by calcium depletion to about 60% (1.8% calcium diet) and 25% (0.38% calcium diet), relative to levels in birds fed the calcium adequate diet (Fig. 4). After repletion, when the ducks were 67 days older, estradiol concentrations in those previously fed calcium-deficient diets were slightly higher (Fig. 4) but not significantly above the control level.

After calcium depletion, the bone breaking strength and mineral density of tibia were decreased with decreasing dietary calcium, but tibia weight and length were not affected. There were no differences in any of these bone characteristics after calcium repletion (Table 4).

The present study confirmed that dietary calcium deficiency impaired eggshell quality by negatively regulating the calcium metabolism, but dietary calcium deficiency did not affect the phosphorus deposition in eggshell. Laying birds obtain the majority of calcium from their diet (60–75%), whereas the remaining 25–40% is taken from skeletal stores (Muller et al., 1964). On calcium-free rations, bone becomes the only source of calcium needed for eggshell. However, it seems impossible for bone to provide the required amount of calcium for eggshell (5–6 g calcium day⁻¹) as supported by the observations that calcium deficiency reduced the plasma concentration of calcium and calcium content in the eggshell, impaired shell microstructure and reduced bone quality. Unexpectedly, there were few shell-less eggs observed during the depletion period because the frequency of ovulation was reduced. This observation is in accordance with previous reports in laying hens that dietary calcium deficiency decreased egg production (Nevalainen, 1969; de Bernard et al., 1980; Abdallah et al., 1993). It probably indicates an adaptable biological physiology of laying birds in response to calcium deficiency. To explore the potential mechanism by which calcium affected eggshell quality, the key proteins involved in the process of shell formation were further assayed.

Osteopontin (OPN, also named SPP1) is a glycosylated, highly phosphorylated protein, expressed in a variety of tissues that are characterized by Ca²⁺ transport, such as bone, kidney and uterus. As one of the major phosphorylated proteins of the avian eggshell matrix (Mann et al., 2007), OPN is synthesized in and secreted from the uterus of the laying bird (Fernandez et al., 2003). OPN localizes specifically in the palisade and mammillae layers (mammillary cores) of the shell (Fernandez et al., 2003; Chien et al., 2008). Its gene expression in the uterus of chicken is stimulated by the presence of the egg in the uterus (Lavelin et al., 2000; Brionne et al., 2014) or due to mechanical distension of the uterine wall (Lavelin et al., 1998). These observations indicate the importance of OPN in shell calcification. Although the mechanism of shell biomineralization is poorly understood, the involvement of OPN in this mechanism may be mediated via the fabrication of OPN-containing fibers (Fernandez et al., 2004; Chien et al., 2008,, 2009) in the collagen type X net of the outer shell membrane that is formed in the isthmus. In the present study, the decreased mammillary knob thickness and mammillae density induced by calcium deficiency may be associated with the suppression of OPN expression in the uterus (SPP1 expression in the duck). This is because the shell calcification takes place upon organic cores associated with fibers of the outer shell membrane, giving rise to the mammillary (inner) and palisade (outer) layers (Sharp and Silyn-Roberts, 1984).

The formation of CaCO₃ is predominantly carried out by the enzyme carbonic anhydrase (CA) because the bicarbonate ions are produced mainly by the glandular cells from metabolic CO₂ in a reaction catalyzed by CA: CO₂+H₂O↔HCO₃⁻+H⁺ (Nys et al., 1999). In this study, the CA-driven deposition of CaCO₃ crystallites in eggshell is largely dependent on the dietary calcium considering that dietary calcium deficiency in laying ducks resulted in decreased deposition of calcium carbonate in the egg shell. This is probably due to the insufficient supply of Ca²⁺ and HCO₃⁻, as well as the reduced transporting capacity of calcium, reflected by the reduced mRNA expression of calbindin, a main calcium-binding protein responsible for calcium transport in the laying bird uterus (Bar, 2009). It is reported that the deposition of calcium into the eggshell of aged laying hens is associated with eggshell gland calbindin (Bar et al., 1992) but the calbindin can be nutritionally manipulated. For example, the activities of calbindin and carbonic anhydrase in the uterus of laying hens were negatively affected after replacing the layer mash with whole grain barley (Balnave et al., 1992), suggesting a potential role of feed in mediating the process of calcium deposition in laying birds.

The eggshell matrix proteins participate in the control of crystal growth through binding and precipitating the crystallized calcium carbonate nanoparticles, therefore regulating the macroscopic properties of the resulting bioceramic. In this study, expression of only two mRNAs encoding eggshell matrix proteins (OC-116 and OCX-32) were determined because the presence of other matrix protein genes (e.g. OC-17) have not been confirmed in ducks. As one of the eggshell matrix proteins, OC-116 has been demonstrated to be the most abundant protein in the identified chicken eggshell proteome (Mann et al., 2007; Mann and Mann, 2013) and the single nucleotide polymorphisms (SNPs) in the OC-116 gene are shown to be significantly associated with the eggshell elastic modulus and thickness, and with egg shape (Dunn et al., 2009). Ovocalyxin-32 (OCX-32) is enriched in the cuticle layer of the eggshell and functions to inhibit the spontaneous precipitation of calcium carbonate during the terminal phase of eggshell formation (Gautron et al., 2001). We show here that dietary calcium deficiency caused reduction in OC-116 and OCX-32 expression in the uterus, accompanied by impaired eggshell structure, although these defects could be reversed by repletion of adequate dietary calcium. The results suggest that the mineralization process of calcium carbonate in the uterus was dependent on the calcium availability in laying birds. Considering the importance of calcium in regulating matrix proteins, the potential molecular mechanism involved needs to be further investigated in vitro. However, the relative mRNA level of clusterin, which functions as a chaperone for the protein matrix (Poon et al., 2000; Mann et al., 2003) was unexpectedly upregulated in the uterus when ducks were fed the calcium-deficient diet in this study. The reason why clusterin expression was upregulated by calcium deficiency remains to be elucidated.

One of the interesting findings in this study is that the estradiol secretion was inhibited by calcium deficiency. Some reports demonstrated the paradoxical role of estradiol in affecting eggshell quality. For example, injecting pellets of estradiol 3-benzoate subcutaneously in laying hens was reported to increase shell thickness and improve shell deformation (Wistedt et al., 2014). In contrast, an injection of 0.5–5 ng g⁻¹ diethylstilbestrol (DES) into the yolk before incubation decreased later egg weight, shell strength and thickness, which accompanied abnormal development of the oviducts after sexual maturation (Kamata et al., 2009). To date, whether the effect of calcium on shell quality is mediated by estradiol remains unclear despite a potential role of estradiol either in mediating calcium metabolism (Bar et al., 1996) or in regulating the key enzyme involved in the process of shell calcification (Bar et al., 1996; Holm et al., 2001; Berg et al., 2004; Kamata et al., 2009).

In conclusion, the present study demonstrates that dietary calcium deficiency impairs the eggshell microstructure, probably by suppressing the process of eggshell formation, including calcium supply and transport, outer membrane formation and calcite crystal calcification, thus negatively affecting eggshell quality.

Animal care procedures outlined by the guidelines of the Animal Care and Use Committee of the Guangdong Academy of Agricultural Sciences were followed for management, housing and slaughter procedures.

At 22 weeks of age, 450 Shanma female ducks (Anas platyrhynchos Linnaeus 1758, a typical breed of laying ducks in South China) were randomly allocated to one of three groups, each with 6 replicate pens of 25 birds. Ducks of one group were fed the control, adequate calcium diet, which contained 3.6% of calcium and two groups were fed calcium-deficient diets, with 1.8% or 0.36% calcium (Table 5). Ducks were maintained on the three diets for 67 days (depletion period), then all ducks were fed the control diet ad libitum for an additional 67 days (repletion period).

On days 67 and 134 (the end of each period), all eggs within each replicate pen (240 or 361 eggs in total for days 67 and 134, respectively) were collected for determination of eggshell quality. At these same times, two ducks within each replicate were randomly selected for blood and tissue sampling. Five ml of blood was collected at 22:00 h from the brachial vein using heparinized tubes (BD vacutainer Systems, Franklin Lakes, NJ), centrifuged (3000 g, 15 min, 4°C) and aliquots of plasma was stored at −80°C. After blood sampling, the ducks were stunned and killed by cervical dislocation at 09:30 h the next day, when approximately 60–70% of the observed eggs entered the uterus. Uterine samples were excised, rinsed briefly with 0.9% isotonic saline and snap-frozen in liquid nitrogen and stored at −80°C. The right and left tibia of each duck were excised from the fresh carcass, individually sealed in plastic bags and stored at −20°C.

Eggs were stored at 4°C overnight and then broken onto a level surface. After weighing the egg mass, the yolk and albumen were separated and the shells were washed under running water, dried and weighed. Shell thickness, without membrane, was measured with a digital micrometer at three locations (air cell, equator and sharp end) and averaged. Shell breaking strength of whole eggs was measured using an Egg Force Reader (ORKA Food Technology Ltd., Ramat Hasharon, Israel).

Two eggs within each replicate pen were sampled and processed for viewing of the ultrastructure of the eggshell. The eggs were broken, the interior contents were removed and the shells were cleaned with tap water. Two shell fragments (1 cm²) were cut from the equator of the shell of each egg, soaked overnight in distilled water. The shell membranes were then carefully peeled from the edge of the sample inward (this step was missed out for cross section analysis). The egg shell samples were then boiled in 2% NaOH for 10 min to dissolve any proteinaceous material incorporated in the shell (Liao et al., 2013). After drying thoroughly at room temperature, the shell samples were then mounted on aluminium stubs using conductive silver paint (1005 aqueous conductive silver liquid-SEM adhesive, ProSciTech, Kirwan, Australia), sputter coated with gold in a Jeol MP-19020NCTR Neocoater and viewed with a XL-30 environmental scanning electron microscope (Philips, Amsterdam, The Netherlands). Photographs of triplicate surfaces were made to facilitate counting the number of mammillary knobs per unit eggshell interior surface area. The density of mammillae of each shell was expressed as the number of knobs per unit (mm²). According to the method of Gongruttananun et al. (2013) and Stefanello et al. (2014), the palisade layer thickness (μm), mammillary knob thickness and total thickness (palisade+mammillae) were assayed from the cross section of shell, with three scanned images obtained for each sample.

Plasma concentration of estradiol was determined with a commercial radioimmunoassay kit (Shanghai Bioproducts, Shanghai, China), following Bluhm et al. (1983) and Yang et al. (2005). The mean within-assay coefficient of variation was 10.2%.

Concentrations of free calcium and phosphorus ions in plasma were determined by an automatic biochemical analyzer (Synchron cx5, Beckman Coulter, Los Angeles, CA, USA), with reagents from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). For assay of calcium and phosphorus in the uterus, ∼0.1 g uterine tissue was homogenized 1:10 (w/w) with ice-cold isotonic saline (0.9%) for 15 s. The homogenates were then centrifuged (3000 g, 15 min, 4°C) and 10 μl of the extract supernatants were used for determination of calcium ion and phosphorus contents with the same kits and expressed as mmol mg⁻¹ protein. The protein content of the extract supernatants was determined with commercial kits (Thermo Fisher Scientific, Waltham, MA, USA).

The eggshells were soaked overnight in distilled water and the membranes were peeled off from the inside of the shells. The shells without membrane were then boiled in 2% NaOH for 10 min, rinsed in distilled water and dried in an oven at 105°C for 2 h. From each shell, two pieces, weighing approximately 450 mg each, were taken at the blunt end of the egg and then heated with 10 ml of 50% 1 N hydrogen chloride (HCl) and five drops of nitric acid until the appearance of white fumes then quantitatively diluted to 100 ml with ion-free water. Calcium was determined by titration with EDTA and UV spectrophotometry. Phosphorus was determined using the colorimetric technique of Goldenberg and Fernandez (1966). Shell calcium and phosphorus were expressed as a percentage of the total shell weight analyzed.

Total RNA was extracted from uterine samples with Trizol (Invitrogen, Carlsbad, CA, USA), treated with DNAase (TAKARA, Dalian, China) and quantified and assessed to be of high quality by OD_260:280 and evaluation after gel electrophoresis. Complementary DNA was prepared using TAKARA RT reagents, according to the manufacturer's instructions. Real-time PCR was performed on 1 μl of the cDNA product in a total volume of 20 μl containing 10 μl of SYBR-green PCR master Mix (TAKARA) and 0.2 μmol l⁻¹ of gene-specific forward and reverse primers (Table 6). The specificity of the reaction was confirmed by determining the product melting curve. The following protocol was used: initial denaturation for 30 s at 95°C, followed by 40 cycles of 20 s at 95°C, 30 s at 60°C and 20 s at 72°C.

The relative expression of each mRNA was calculated by the ΔC_t method as described previously (Livak and Schmittgen, 2001). This is the value obtained by subtracting the C_t value of β-actin (ACTIN) mRNA from the C_t value of the target mRNA, and it was then used to normalize the relative expression level of the target gene [2^{−(ΔCt treatment−ΔCt control)}, or 2^−ΔΔCt].

The frozen tibias were thawed inside the plastic bags at room temperature for 2 h and tibial weight and length were measured on two tibias per replicate. Bone mineral density (BMD) and breaking strength were subsequently determined as described in detail previously (Chen et al., 2015).

Replicate (N=6) was taken as the experimental unit. The effect of treatment was assessed by one-way ANOVA using the GLM procedure of SAS8.0 (SAS Institute, Cary, NC, USA). Where appropriate, differences in variables between treatments were compared by Tukey tests. Students' t-test was used to compare the differences in expression of gene transcripts and plasma levels of estradiol between the calcium-deficient groups and calcium-adequate controls. Significance was established at P<0.05.

We sincerely thank Dr W. Bruce Currie (Emeritus Professor, Cornell University) for his help in presentation of this manuscript.