Impaired Bone Formation in Pdia3 Deficient Mice

1α,25-dihydroxyvitamin D3 [1α,25(OH)2D3] is crucial for normal skeletal development and bone homeostasis. Protein disulfide isomerase family A, member 3 (PDIA3) mediates 1α,25(OH)2D3 initiated-rapid membrane signaling in several cell types. To understand its role in regulating skeletal development, we generated Pdia3-deficient mice and examined the physiologic consequence of Pdia3-disruption in embryos and Pdia3 +/− heterozygotes at different ages. No mice homozygous for the Pdia3-deletion were found at birth nor were there embryos after E12.5, indicating that targeted disruption of the Pdia3 gene resulted in early embryonic lethality. Pdia3-deficiency also resulted in skeletal manifestations as revealed by µCT analysis of the tibias. In comparison to wild type mice, Pdia3 heterozygous mice displayed expanded growth plates associated with decreased tether formation. Histomorphometry also showed that the hypertrophic zone in Pdia3 +/− mice was more cellular than seen in wild type growth plates. Metaphyseal trabecular bone in Pdia3 +/− mice exhibited an age-dependent phenotype with lower BV/TV and trabecular numbers, which was most pronounced at 15 weeks of age. Bone marrow cells from Pdia3 +/− mice exhibited impaired osteoblastic differentiation, based on reduced expression of osteoblast markers and mineral deposition compared to cells from wild type animals. Collectively, our findings provide in vivo evidence that PDIA3 is essential for normal skeletal development. The fact that the Pdia3 +/− heterozygous mice share a similar growth plate and bone phenotype to nVdr knockout mice, suggests that PDIA3-mediated rapid membrane signaling might be an alternative mechanism responsible for 1α,25(OH)2D3’s actions in regulating skeletal development.

PDIA3 is a multiple function protein of the protein disulfide isomerase family [6]. It is well known as a chaperonin, which can assist glycoprotein folding and major histocompatibility complex I loading [7]. It regulates endoplasmic reticulum stress and controls cell survival [8]. It has also been shown to mediate 1a,25(OH) 2 D 3 initiated rapid membrane responses in a variety of cell types. Nemere et al. have demonstrated that PDIA3 directly contributes to 1a,25(OH) 2 D 3 induced calcium uptake by intestinal cells in chicks and rats, which is important for mineral ion homeostasis and bone mineral mass accumulation [9].

D 3 regulation of skeletal development in vivo.
Cell culture studies support this. Growth plate chondrocytes from nVdr knockout mice continue to exhibit increased PLA2, PLC and PKC activity in response to 1a,25(OH) 2 D 3 [14]. The mechanism by which 1a,25(OH) 2 D 3 elicits its actions in growth plate chondrocytes involves direct interaction with caveolin-1 (CAV-1) in caveolae microdomains [15]. Interestingly, growth plate chondrocytes from Cav-1 knockout mice fail to respond to 1a,25(OH) 2 D 3 with an increase in PKC [16]. Similarly, chondrocytes lacking intact caveolae also lack this rapid response to the vitamin D metabolite. Moreover, Cav-1 2/2 knockout mice exhibit a growth plate phenotype characterized by few hypertrophic chondrocytes compared to their Cav-1 +/+ littermates [15].
The skeletal actions of 1a,25(OH) 2 D 3 have been studied in animal models and clinically. Vitamin D deficiency in animals and children results in rickets, characterized by inadequate calcification of the growth plate and adjacent metaphyseal bone [1,17]. Several nVdr knockout mice models have been created to study the mechanisms involved in the development of the skeletal phenotype of hereditary vitamin D resistant rickets, a human disease associated with nVdr mutations [18][19][20]. These mice exhibit a rachitic growth plate phenotype after weaning, supporting the hypothesis that a functional VDR is required for bone development. Later studies showed that the rickets could be healed by a calcium-rich rescue diet, suggesting that the rachitic changes are a direct consequence of impaired endocrine function of 1a,25(OH) 2 D 3 on mineral homeostasis, rather than impaired VDR-dependent genomic actions on bone [21].
Recent studies examining effects of the rescue diet on the morphology of the growth plate in nVdr 2/2 mice showed that although the rescue diet restored mineral deposition, it did not restore formation of mineralized tethers to normal levels [22]. Tethers are mineralized regions in the growth plate that link the epiphysis to the metaphysis, providing biomechanical stability during long bone growth. They are distributed in a distinctive pattern inside the circumference of normal growth plates as well as within the central region. nVdr 2/2 mice lack the circumferential tethers whether they are fed the rescue diet or not [22], indicating that their formation is VDR-dependent and is not secondary to 1a,25(OH) 2 D 3 's actions on mineral ion homeostasis. Whether they are PDIA3-dependent or not is not known.
Although these in vitro and in vivo data collectively suggest an important role for PDIA3-mediated 1a,25(OH) 2 D 3 rapid membrane signaling in skeletal development, its involvement in endochondral bone formation in vivo remains unknown. Attempts to generate a Pdia3-null mouse have met with frustration, as global ablation of the protein has resulted in embryonic lethality [7,23]. Accordingly, mice that exhibited specific knockout in Bcells were established to assess the role of PDIA3 in immune function [7] and mice with specific knockout in intestinal epithelial cells were established to assess the role of PDIA3 in mineral ion transport [23]. In the present study, we opted to retain the global knockout approach and to examine heterozygous mice in order to gain a better understanding of how PDIA3 contributes to post-fetal musculoskeletal development, which involves multiple tissues and both systemic and biomechanical regulatory stimuli. Our results show that Pdia3 +/2 mice exhibit a mild skeletal phenotype characterized by enlarged growth plates, reduced tether and trabecular bone formation, and impaired osteogenic potential of calvarial osteoblasts in vitro.

Ethics statement
All animal work was conducted under the approval of the Georgia Institute of Technology IACUC.
The ESCs were injected into C57BL/6 blastocysts to create chimeric mice, which were bred with C57BL/6 mice to generate heterozygous Pdia3 +/2 deficient mice. The Pdia3 +/2 offspring were then intercrossed. Mice were genotyped by PCR amplification of tail DNA using prime pairs of a-b and a-c (Fig. 1A). Sequences of the primer sets used for PCR are listed in Table S1. The mice were weaned at 21 days of age, housed in a barrier facility with a 12-h light-dark cycle, and maintained on standard chow.

Examination of mice embryos
To obtain embryos from different stages, Pdia3 +/2 heterozygotes were intercrossed to produce timed pregnancies. At 1 p.m., we checked the vaginal plug and considered it as E0.5. Embryos were isolated at different days of the pregnancy and stained with arcridine orange (Sigma, USA) in PBS for 30 min at 37uC. After extensive washes with PBS, the embryos were examined by stereoconfocal microscopy. For detection of transgenic b-galactosidase activity, the embryos were subjected to X-gal (Sigma, USA) staining according to the manufacturer's protocol. After examining the embryos microscopically, RNA, DNA and protein were extracted from them using Trizol according to the manufacturer's instruction (Invitrogen, USA). DNA was subjected to genotyping using primers described above, RNA was reverse transcribed into cDNA and amplified using primers against Pdia3, ribosomal protein s18 (S18) and neomycion (Neo). Sequences of the primer sets used for PCR are listed in Table S1.

Gross examination of adult mice
The wild type and heterozygous mice at 5, 10, 15 and 30-weeksold were euthanized with CO 2 inhalation following the IACUC protocol. The body length was measured as the distance between the nose and the end of the tail using a conventional ruler and the bodies were weighed. Multiple organs including liver, kidney, and heart were carefully dissected and weighed. The tibias on both sides were dissected, the surrounding skin and muscle were removed and the length of the tibias was measured using a caliper. Expression of Pdia3 in the mice was measured in liver tissue to verify that the heterozygotes had reduced levels of this mRNA.

mCT imaging and analysis
High-resolution mCT images of the left tibia of 5, 10, 15 and 30week-old Pdia3 +/+ and Pdia3 +/2 mice (n = 6) were obtained using a Scanco VivaCT40 (Scanco Medical, Basserdorf, Switzerland) with a voxel resolution of 12 mm. Specific regions of analysis included the proximal tibial epiphysis, metaphysis, mid-diaphysis and distal growth plate. Standard Scanco software was used to determine microarchitecture morphology [16]. The growth plate was isolated with user-guided contours and evaluated at a threshold corresponding to 250 mg hydroxyapatite/cm 3 . Parameters derived at the tibial epiphysis and metaphysis included tissue volume (mm 3 ), bone volume (mm 3 ), bone volume/tissue volume (BV/TV %), trabecular number (#/mm), thickness (mm), and spacing (mm). At the cortical tibia mid-diaphysis, mean cortical area (mm 2 ) and cortical thickness were determined through the scanned volume of bone.
Tethers are bony structures in the growth plate linking the epiphysis and metaphysis and are present in mice, rats, dogs, pigs, and humans [22,28]. They have been linked to changing growth plate morphology and fusion, and to providing tissue stability. Our previous study extensively characterized the tether structure in mouse growth plates and demonstrated its correlation with functional nVdr [22,28]. Therefore, in this study, we adopted this parameter to characterize Pdia3 heterozygous growth plates in addition to other conventional measurements including growth plate cartilage volume and properties of the hypertrophic zone. To qualitatively assess tissue morphology and tether distribution, 3D reconstructions of the growth plate were used to create 3D color images to visualize growth plate morphology and 3D tether maps to visualize formation within the growth center and along the surface [28]. Parameters derived in growth plate including growth plate volume (mm 3 ) and tether volume/growth center volume.

Histology and histomorphometric analysis
For histology and histomorphometric analysis, the left tibiae were fixed overnight in 4% paraformaldehyde. After rinsing in PBS, the samples were subjected to decalcification by incubating with cal-EX solution (Fisher Scientific Inc., USA) for 2 weeks. Decalcified samples were then embedded in paraffin and sectioned at 5-mm thickness. After deparaffinization, the sections were then subjected to trichrome staining. Images were captured using a Nikon Eclipse E100 microscope (Nikon Instruments Inc., Melville, NY). The thickness of the proximal tibial growth plate and the number of hypertrophic cells were then measured using Image Pro software. In vitro osteogenic differentiation of bone marrow progenitor cells isolated from wild type and heterozygous mice Phenotypic changes were evident in the tibias at 15 weeks of age (see Results below), suggesting that changes at the cellular level had occurred before that time. Accordingly, bone marrow stromal cells (BMSCs) were isolated from 10-week-old wild type and heterozygous mice tibiae as described previously [16]. Briefly, tibias from Pdia3 +/+ and Pdia3 +/2 mice (n = 6 per group) were flushed using a 21-gauge needle attached to a 10 mL syringe filled with DMEM (GIBCO BRL, Gaithersburg, MD). Cells were filtered through a cell strainer with 70-micron nylon mesh (BD Bioscience, Bedford, MA) and then one million cells were plated in each well of a 12-well plate and cultured in mesenchymal stem cell growth media (a-MEM containing 10% FBS) at 37uC with 5% humidified CO 2 . The media were changed after 24 hours to remove non-attached cells and replaced with osteogenic medium (a-MEM containing 10% FBS, 0.2 mM dexamethasone, 10 mM b-glycerol phosphate and 50 mg/mL ascorbic acid). Media were replaced every 2 days for up to 28 days of culture. For measurement of osteoblast gene expression, cultures were harvested at day 14 and day 28. RNA was extracted using TRIzol (Invitrogen) and reverse-transcribed into cDNA using the Omniscript RT kit (Qiagen, Valencia, CA) according to the manufacturer's directions. Real-time PCR was performed using SYBR Green SuperMix 170-8882 (Bio-Rad) for collagen type I (Col1), alkaline phosphatase (Alp), Runt-related transcription factor 2 (RunX2), bone sialoprotein (Bsp), osteopontin (Opn), and osteoprotegerin (Opg) and normalized to the level of housekeeping gene ribosomal protein s 18 (S18). The real-time PCR primers are listed in Table S2. Alkaline phosphatase activity was measured at day 28 and normalized to protein content [12]. The level of osteocalcin in the media at day 28 was measured using a mouse osteocalcin ELISA kit (Biomedical Technologies, Inc.) according to manufacturer's protocol and normalized to DNA content. To assess mineralization, the 28-day cultures were fixed in 10% formalin; alizarin red and von Kossa staining were performed, and the positive staining area was quantified using ImageJ software [16].

Statistical analysis
All measurements were collected with N = 6 independent determinations per data point. The data were expressed as mean 6 standard error and were subjected to independent Student's t tests. All statistical analyses were carried out using GraphPad 5.0 (GraphPad Software, Inc., La Jolla, CA). P,0.05 were considered as statistically significant.

Generation of pdia3 deficient mice
The functional disruption of the Pdia3 gene was confirmed by the loss of response to 1a,25(OH) 2 D 3 induced rapid PKC activation in Pdia3 2/2 RST613 ESCs. While 1a,25(OH) 2 D 3 treatment of wild type ESCs for 9 min resulted in a 4.9 fold increase in PKC activity compared to vehicle treatment (P,0.05), no significant PKC activation was observed in Pdia3 2/2 RST613 ESCs (Fig. 1B). These RST613 ESCs were then used to generate Pdia3 knockout mice.
Global disruption of the Pdia3 gene resulted in embryonic lethality. To determine the time of embryonic lethality, embryos were collected from heterozygous intercrosses at different days of gestation. No Pdia3 2/2 embryos were found after embryonic day 12.5. At E10.5, 2 out of 25 embryos were detected as Pdia3 2/2 by PCR analysis (Fig. 1C). The absence of Pdia3 mRNA (Fig. 1D) and protein (Fig. 1E), and the presence of b-galactosidase activity (Fig. 1F) were confirmed by RT-PCR, western blot, and X-gal staining, respectively.
Heterozygous mice were fertile, exhibited a normal life span and did not show prominent pathological alterations. Genotype analysis of 500 offspring from Pdia3 +/2 intercrosses revealed 33% were Pdia3 +/+ wild type, 67% were Pdia3 +/2 heterozygotes, and none were Pdia3 2/2 homozygous. Gross anatomy analysis (Fig. 3) showed that Pdia3 +/2 mice had short body lengths (week 10, 15 and 30) but long tibias (week 15), and larger heart/body ratios (week 15 and 30) in comparison to wild type mice. No significant differences in body weight, body weight/body length, and kidney/body ratio were observed.
Expression of 1,25(OH)2D3 receptors in pdia3 +/2 mice during the course of development Real-time PCR analysis of liver tissue (Fig. 4) showed that the expression of Pdia3 in heterozygous mice was 30-50% lower than that of wild type mice through week 15 but by week 30, differences in Pdia3 expression between wild type mice and heterozygous mice were no longer evident. In contrast, nVdr expression in the liver was similar between wild type and heterozygous mice at all observation time points. nVdr expression increased at week 10 then decreased to the basal level (P,0.05) in both groups of mice.
Growth plate abnormality in pdia3 +/2 mice The growth plates of 10-week-old wild type mice contained a number of tethers that displayed as bony bridge structures crossing the growth plate and connecting the epiphysis and metaphysis in 2D mCT images (Fig. 5A) or hole-like structures in 3D reconstruction images (Fig. 5B). However, these structures were barely observed in the growth plates of heterozygous mice (Fig. 5C, D). Quantitative analysis showed an age-dependent increase in tether volume in both types of animals but it was significantly lower in Pdia3 +/2 mice in comparison to wild type mice at week 5, 10 and 15 (Fig. 5F). In contrast, the cartilage volume of growth plates was significantly higher in Pdia3 +/2 mice compared to wild type animals at both week 10 and 15 (Fig. 5E). Histomorphometric analysis of histological sections also confirmed that the growth plates of 10-week-old heterozygotes were about 20% thicker and hypertrophic zones contained 12% more cells than those of wild type mice (Fig. 5G-L). Taken together, these results consistently and collectively suggested a decrease of mineralization of growth plate in Pdia3 +/2 mice.
Bone phenotype in pdia3 +/2 mice An age-dependent bone phenotype was observed in tibias of Pdia3 +/2 mice by mCT and morphometric analysis (Tables 1 and  2). At week 5, no significant difference in trabecular bone was found between wild type and heterozygous mice but at week 10, Pdia3 +/2 mice had a significantly lower number of trabeculae within the epiphysis and BV/TV within the metaphysis in comparison to wild type mice. At week 15, Pdia3 +/2 mice displayed significantly lower BV/TV and smaller trabecular thickness, but larger trabecular space within both the epiphysis and metaphysis compared to wild type. At week 30, no significant difference was observed within the epiphysis. In contrast, the Pdia3 +/2 mice exhibited significantly higher BV/TV within the metaphysis in comparison to wild type. Notably, no significant difference in cortical bone was found at all ages (data not shown).
Impaired osteoblast differentiation potential of BMSCs isolated from pdia3 +/2 mice Osteoblast differentiation of BMSCs was affected by Pdia3 deficiency. Although the expression of col1 and runx2 was similar between the two types of cultures (P.0.05), the expression of several osteoblast markers including alp, bsp, opn and opg was significantly lower in Pdia3 +/2 cultures in comparison to wild type cultures at day 14 or day 28 (Fig. 6 A-F). In addition, heterozygous cells also displayed significantly lower alkaline phosphatase activity and produced less amounts of osteocalcin than wild type cells (Fig. 6 G, H). By the end of 28 days of osteogenic differentiation, the deposition of mineralized extracellular matrix was less in heterozygous cultures than in wild type cultures as indicated by smaller positive staining area of Alizarin red and von Kossa (Fig. 6I).

Discussion
Due to embryonic lethality of global knockout of Pdia3, we developed mice with null mutation in one allele of the gene in order to study its role in skeletal development in post-fetal mice. Comparison of wild type and heterozygote mice showed that Pdia3 deficiency results in mild skeletal manifestations at both growth plate and trabecular bone levels. The expansion of the growth plate was associated with an increased hypertrophic zone but decreased tether formation, which might contribute to the reduced trabecular bone volume [15,28]. The impaired bone formation can also be attributed to reduced osteogenic potential of bone marrow cells in Pdia3 +/2 mice as evidenced by their lower mineralization capacity and reduced expression levels of osteoblast markers when cultured in osteogenic media. Changes in bone quantity may also be caused by altered bone resorption [16], which was not investigated in this study.
The skeletal manifestations in Pdia3 +/2 mice varied with postnatal age and on the skeletal region. Lower numbers of growth plate tethers in Pdia3 +/2 mice were first detected at week 5. The reduced number of tethers did not change with time, but growth plate volume was larger, the number of hypertrophic cells was greater and the number of trabeculae was reduced at week 10 in the Pdia3 +/2 mice. The differences in bone parameters between wild type mice and heterozygotes were most pronounced at week 15 in both epiphysis and metaphysis. To our surprise, while no significant differences were found in the growth plate and epiphyseal bone at week 30, Pdia3 +/2 mice displayed a higher metaphyseal bone volume in comparison to wild type mice. This observation was consistent with our previous study on the femoral metaphysis of 15-week-old Pdia3 +/2 mice, which showed significant increases in trabecular and cortical bone volume [26].
The mechanisms underlying the age-and location-dependent skeletal phenotypes of Pdia3 +/2 mice are still not clear. However, recent studies provide some clues that may contribute to the observations reported here. First, Pdia3 and nVDR might serve as a dominant 1a,25(OH) 2 D 3 signaling mediator at different development stages. Others have shown that in chickens, 1a,25(OH) 2 D 3 elicits rapid stimulation of phosphate and calcium transport in paralle with PKC activation in young birds but not adult birds, due to gradual decrease in the binding affinity of PDIA3 for 1,25(OH) 2 D 3 [9]. In contrast, nVDR signaling remains either unchanged or increases with age [9]. We reported that PDIA3 is present in ESCs and mediates rapid 1a,25(OH) 2 D 3dependent PKC activation prior to expression of nVdr [29]. Moreover, Pdia3 expression remains constant, whereas nVdr expression increases over time. Similarly, in the present study, Pdia3 expression was constant over time but there was an increase in nVdr expression at week 10. The reduction in nVdr mRNA at later time points suggests that PDIA3 signaling may play a more dominant role in mediating the effects of 1a,25(OH) 2 D 3 after week 10, which might explain why disruption of Pdia3 led to the majority of bone phenotypes to occur in heterozygotes at week 15. Further examination of PDIA3 protein production in the growth plate and bone as well as functional disruption of 1a,25(OH) 2 D 3 induced PKC activity in these Pdia3 +/2 cells are essential to elucidate the cellular mechanism of PDIA3's action on skeletal development.  The differential influence of disruption of Pdia3 on bones at different regions might also be due to their unique skeletal development process, geometrical and mechanical properties. The epiphysis is formed during embryonic bone formation, whereas the metaphysis is established post-fetally [16]. A progressive skeletal phenotype was also observed in Cav-1 2/2 mice from epiphysis to metaphysis [16]. Compared to the epiphysis, metaphyseal bone also has much larger contact surface area with bone marrow, which can regulate the bone formation through dynamic remodeling [30]. Moreover, femurs and tibias have distinct geometrical and mechanical properties [31,32], which can also lead to their different phenotypic responses to PDIA3's actions.
The skeletal phenotype of Pdia3 +/2 mice was similar but not identical to those of nVdr 2/2 mice and/or Cav-1 2/2 mice. Additionally, the extent of skeletal manifestations in Pdia3 +/2 mice was not as pronounced as in the other two models, which might be due to the partial deletion of Pdia3 rather than complete ablation of nVdr or Cav-1, respectively. All of these animals have expanded hypertrophic cartilage. Both nVdr 2/2 and Pdia3 +/2 mice have reduced tether formation indicating of failure of calcification of cartilage [22,28], whereas a delayed closure of the growth plate in Cav-1 2/2 mice is responsible for growth plate expansion [15,16]. Both nVdr 2/2 and Pdia3 +/2 mice have reduced trabecular bone volume whereas Cav-1 2/2 mice exhibited increased bone formation [16,22]. In terms of osteogenic potential, both osteoblasts from nVdr 2/2 mice and bone marrow cells from Cav-1 2/2 mice have demonstrated enhanced mineralized matrix formation and increased expression of osteoblast markers [16,33]. In contrast, osteoprogenitor cells from bone marrow of Pdia3 +/2 mice showed decreased osteogenic potential.
While it is not surprising that these three mice models share some features in their skeletal phenotypes in view that nVDR, PDIA3, and CAV-1 are all involved in 1a,25(OH) 2 D 3 signaling pathways and the potential cross-talk among them, their unique features indicate that these proteins might regulate skeletal development through mechanisms other than as 1a,25(OH) 2 D 3 effectors. It has been suggested that CAV-1 can recruit b-catenin to caveolae and facilitate its subsequent nuclear translocation to activate the canonical Wnt pathway and regulate osteoblast differentiation and stem cell renewal [34,35]. Our in vitro studies also show that silencing nVdr or Pdia3 in osteoblast-like MC3T3-E1 cells leads to increased expression of BMP-2 and its inhibitor Noggin [13], which indicates that these two proteins might be  involved in osteoblast commitment thorough regulation of the BMP-2 signaling pathway. Of note, in addition to their skeletal phenotype, a greater heart size of Pdia3 +/2 mice was also observed after week 15, which indicates the potential role of PDIA3 in cardiovascular development. Interestingly, nVdr knockout and heterozygous mice also exhibit larger hearts than wild type mice as well as cellular hypertrophic changes in heart tissue at 12-months, which is directly due to ablation of nVDR signaling [36,37]. Further studies of the cardiovascular system in PDIA3-deficient mice will be needed in order to reveal the mechanism involved.
In this study, no Pdia3 knockout mice were obtained after birth confirming that global Pdia3 knockout is embryonically lethal, as has been reported by others using different strategies for generating the null mutation [7,38]. The Pdia3 2/2 embryo ratio (8%, 2 of 25) at E10.5 was far below the predicted ratio of 25% based on Mendelian distribution. The underrepresented Pdia3 2/2 embryos at this stage indicated the reabsorption of mal-developed Pdia3 null embryos in the uterus and suggested that Pdia3 2/2 embryos died before E10.5. Although the two identified null embryos were morphologically abnormal, the exact onset and causes of embryonic lethality were not defined in this study. Others have reported that Pdia3 ablation caused embryonic lethality at E13.5 [38]. Different from our knockout strategy with insertion of the gene trap vector in intron 1 of the Pdia3 alleles, that study inserted the vector into intron 4 and obtained a longer truncated PDIA3 protein containing partial functional groups [38], which might explain the different embryonic lethality stages observed between these two studies.
In conclusion, our results provide for the first time the in vivo evidence that PDIA3 is critical for normal skeletal development in post-natal mice. Partial disruption of Pdia3 leads to impaired bone formation and genetic ablation of Pdia3 causes embryonic lethality. Because the Pdia3-deficiency in this animal model is not limited to cartilage or bone cells, it is not possible to define the exact mechanism by which PDIA3 controls skeletal development. In vitro studies examining the signaling pathways involved in PDIA3-dependent effects of 1a,25(OH) 2 D 3 in growth plate chondrocytes and osteoblasts [39] support the hypothesis that downstream actions of the vitamin D metabolite on cell function are responsible for some of the observed phenotypic outcomes in the Pdia3 +/2 mice. Development of conditional knockout mice with Pdia3 inactivation in cartilage or bone is essential to elucidate the direct role of PDIA3 in endochondral bone formation. Chondrogenesis precedes osteogenesis during embryonic development; thus, a bone- Figure 6. Reduced osteogenic differentiation capacity of bone marrow stem cells isolated from pdia3 +/2 mice. The bone marrow stem cells (BMSCs) were isolated from wild type and heterozygous mice and cultured in osteogenic media for 14 and 28 days. The expression of collagen type I (Col1, A), alkaline phosphatase (Alp, B), Runt-related transcription factor 2 (Runx2, C), bone sialoprotein (Bsp, D), osteopontin (Opn, E) and osteoprotegerin (Opg, F) were evaluated by real-time PCR and normalized to housekeeping gene ribosomal protein s 18 (S18). Alkaline phosphatase activity (G) and the production of Osteocalcin (H) were measured in Day 28 cultures. Alizarin red and von Kossa staining were performed on Day 28 cultures to assess the mineralization and the positive staining areas were measured (I). White bar represents wild type culture and black bar represents heterozygous culture. N = 6, *P,0.05 vs. wild type. doi:10.1371/journal.pone.0112708.g006 specific conditional knockout would not necessarily capture effects of Pdia3 ablation on musculoskeletal outcomes. Moreover, endocrine actions will not be excluded from these conditional knockout systems. Mineral homeostasis is affected in chondrocytespecific nVdr-null mice [40], but it is not known whether PDIA3deficiency will also impact mineral metabolism. Therefore, levels of Ca ++ and phosphate in the blood as well as analysis of hormones or other factors involved in bone or mineral homeostasis will be vital to address the systemic effects of PDIA3 in controlling skeletal development [41]. Additionally, in vitro studies using Pdia3 knockout ESCs will be a useful model at cellular level for investigating the engagement of PDIA3-mediated 1,25(OH) 2 D 3 signaling and its potential crosstalk with nVDR and CAV-1 signaling pathways in skeletal development. Finally, given the fact that PDIA3 can regulate skeletal development, it might serve as a target for musculoskeletal tissue engineering applications in the future.