Wntless spatially regulates bone development through β-catenin-dependent and independent mechanisms

Abstract

Background

Canonical and noncanonical Wnt signaling pathways both play pivotal roles in bone development. Wntless/GPR177 is a chaperone protein that is required for secretion of all Wnt ligands. We previously showed that deletion of Wntless within mature osteoblasts severely impaired postnatal bone homeostasis.

Results

In this study we systemically evaluated how deletion of Wntless in different stages of osteochondral differentiation affected embryonic bone development, by crossing Wntless (Wls)-flox/flox mice with strains expressing cre recombinase behind the following promoters: Osteocalcin, Collagen 2α1, or Dermo1. Ex vivo μCT and whole-mount skeletal staining were performed to examine skeletal mineralization. Histology and immunohistochemistry were utilized to evaluate cellular differentiation and alterations in Wnt signaling. In this work, we found that Wntless regulated chondrogenesis and osteogenesis through both canonical and noncanonical Wnt signaling.

Conclusion

These findings provide more insight into the requirements of different Wnt-secretion cell types critical for skeletal development.

Introduction

Wnt signaling has been intensively studied for over three decades [1, 2]. Our understanding of Wnt signaling in bone has increased exponentially since the identification of mutations in the Wnt co-receptor LRP5 as the underlying cause of the osteoporosis pseudoglioma (OPPG) syndrome [3]. Both β-catenin-dependent (or canonical) and –independent (or noncanonical) Wnt signaling pathways play important roles in prenatal and postnatal bone development; many mouse models with deficiencies in Wnt components have been characterized for skeletal phenotypes [4, 5]. However, relatively fewer studies have focused on the source of Wnt ligands relevant for normal skeletal development. There are 19 Wnt proteins in the mammalian genome and there is abundant in vivo and in vitro evidence suggesting multiple Wnt ligands are involved in skeletal development in β-catenin-dependent and -independent manners [6–11]. Among all reported Wnt knockout mice, the Wnt3a global knockout (Wnt3a−/−) embryo displayed the earliest skeletal abnormalities with a defect in vertebrate axis formation [12]. Wnt9a−/− mice developed a suture fusion defect in the skull, while Wnt9a−/−;Wnt4−/− double mutants showed more severe phenotypes [13]. More recently, WNT1 heterozygous mutations in humans were shown to be associated with early-onset osteoporosis in osteogenesis imperfect (OI), and the Swaying (Wnt1^sw/sw) mouse model carrying a spontaneous mutation in Wnt1 also showed phenotypes consistent with OI [14, 15]. Many other global Wnt knockout mouse models have been developed and some display skeletal phenotypes [9, 11, 16–18]. However, in many cases, the phenotypes were either too profound outside of skeletal system or interpretation was complicated by the compensatory roles from other Wnt ligands.

Wntless (WLS) was identified as an exclusive chaperone protein for Wnt, and is required for secretion of all mammalian Wnts [19, 20]. Genome-wide association study (GWAS) studies also showed that SNPs in the human WLS locus (also known as GPR177) was highly associated with changes in human bone mineral density [21]. Germline homozygous deletion of Wls in mice causes early embryonic lethality associated with dramatic body axis defects, while mesenchymal- or ectoderm-specific deletion of Wls caused defects in limbs and other tissues [22, 23]. We and others previously characterized osteoblast-specific Wls knockouts and found that mutant animals displayed a dramatic osteoporotic phenotype [24, 25]. Confirming the importance of osteoblast-derived Wnts for postnatal skeleton, an osteoblast-specific Wnt16 knockout mouse model displayed a cortical bone phenotype with increased fractures [26]. In this work, we generated three conditional deletions of Wls within the osteochondral lineage using Osteocalcin (Ocn)-Cre (Tg(BGLAP-cre)1Clem) [27], Col2a1-Cre (B6;SJL-Tg(Col2a1-cre)1Bhr/J) [28], and Dermo1-Cre (Twist2^tm1(cre)Dor) [29] promoters to determine the requirements for different cellular sources of Wnt ligands in embryonic bone development.

Results and Discussion

1. Characterization of bone-specific promoters

To confirm the specificity and penetrance of cre expression relevant for these studies, we crossed the well-established mT/mG reporter mouse strain (Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J) [30] to each of the three cre strains used: Osteocalcin (Ocn) (Tg(BGLAP-cre)1Clem) [27], type II collagen (Col2) (B6;SJL-Tg(Col2a1-cre)1Bhr/J) [28], and Dermo1 (Twist2^tm1(cre)Dor) [29] promoters. In the mT/mG mouse strain (B6.129(Cg)-Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J) [30], cells constitutively express the tomato protein (red) in cells unless they have been exposed to cre recombinase during some point of their development. Exposure to cre recombinase results in deletion of the tomato cassette and the induction of GFP expression. Thus, all cells descended from a cell that was exposed to cre recombinase can be identified by the presence of green fluorescence. We used a stereo fluorescent microscope to characterize green fluorescence (Cre activity) of Ocn-Cre;mT/mG, Col2a1-Cre;mT/mG, and Dermo1-cre;mT/mG embryos. Ocn-Cre;mT/mG embryos showed GFP expression as early as E14.5 in some skeletal components, such as skull, jaw, and joints (Fig. 1Aa). In contrast, Col2a1-Cre;mT/mG showed very strong GFP expressions in most skeletal components except the frontal bone of calvarium at E14.5 (Fig. 1Ba). The E14.5 Dermo1-Cre;mT/mG embryos showed GFP expression throughout the body, so that the green fluorescence within the skeleton could not be distinguished in the stereo images (Fig. 1Ca).

Figure 1. Specificity of bone-specific promoters.

mT/mG reporter mice were crossed with Ocn-Cre, Col2a1-Cre or Dermo1-Cre transgenic mice. Whole-mount fluorescence of E14.5 embryos (carrying both Cre and mT/mG transgenes) was evaluated with a fluorescent Olympus MVX10 stereomicroscope (Aa, Ba and Ca). The green fluorescence (GFP) marks the tissue with Cre activity. Cryosections of distal femurs from E17.5 embryos were observed with a fluorescent microscope, and the corresponding bright-field pictures are also shown (Ab, Bb and Cb). Distal femur and femoral middle shaft from 1-month-old Ocn-Cre;mT/mG mice were sectioned and subjected to anti-GFP IHC, arrows indicate osteocytes with positive staining and arrow heads indicate osteocytes with negative staining (Ad). A fluorescent image of a typical mT/mG distal femur (E17.5, no Cre) is shown (D). Please note that there is extremely low background green fluorescence observed under the same imaging conditions. Growth plate regions (GP) are indicated. Yellow arrows indicate the bone-surrounding cells expressing GFP in Dermo1-Cre;mT/mG. The scale bars indicate 400 μm, or otherwise indicated.

To observe evidence for Cre activity in more detail, we collected E17.5 mT/mG embryos which had been intercrossed with the Cre strains used for this work, and sectioned the distal femurs to carefully assess green fluorescence. In Ocn-Cre;mT/mG, we observed that the majority of Cre activity was within osteoblasts in trabecular bone and endosteum at E17.5 (Fig. 1Ab), In addition, most osteocytes within the femoral cortex of 1-month-old mice also displayed green fluorescence (Fig. 1Ad). In contrast, Col2a1-Cre;mT/mG showed GFP expression in both cartilage and bone of distal femur (Fig. 1Bb). This observation is in line with several reports that Col2a1-Cre probably targets a population of osteochondral progenitor cells that differentiate into osteoblasts [31–36]. Consistent with the stereo fluorescent images, GFP expression in Dermo1-Cre;mT/mG showed a wide spread pattern in distal femur including bone-surrounding tissues (Fig. 1Cb, arrows).

2. Wnts from different cellular sources differentially regulate embryonic skeletal mineralization

Using a mouse strain with a conditionally inactivatable allele of Wls (129S-Wls^tm1.1Lan/J), we previously characterized Ocn-Cre;Wls^flox/flox mice, and observed a dramatic osteoporotic phenotype in young and postnatal mutant animals, indicating that Wnts from mature osteoblasts are important for postnatal bone homeostasis [24]. In this study we sought to examine embryonic skeletal development by blocking Wnt secretion from cells derived from Ocn-cre, Col2a1-Cre, or Dermo1-Cre positive cells by crossing these strains to the Wls^flox allele.

We did not observe any gross defects in the skeleton of Ocn-Cre;Wls^flox/flox mutant embryos up to E18.5 using μCT scanning (Fig. 2Ab) or whole mount staining (Fig. 2Ad). However, we did observe less bone mass in the skulls (but not long bones) of 5-day-old mutant Ocn-Cre;Wls^flox/flox pups (Fig. 3Bb).

Figure 2. Wntless deletion causes delayed embryonic skeletal mineralization.

MicroCT images were reconstructed into 3-D models to visualize the skeleton of E18.5 Ocn-Cre; Wls^flox/flox (Aa and Ab) and E16.5 Col2a1-Cre;Wls^flox/flox (Ba and Bb) embryos. Whole-mount skeletal staining was performed on E18.5 Ocn-Cre;Wls^flox/flox (Ac and Ad), E16.5 Col2a1-Cre; Wls^flox/flox (Bc and Bd), and E14.5 Dermo1-Cre;Wls^flox/flox (Ca and Cb) embryos. The arrows indicate the regions with less mineralization in mutant embryos. Note there is little mineralization in mutant Dermo1-Cre;Wls^flox/flox embryos (Cb).

Figure 3. Wntless deletion causes defects in skull and long bone mineralization.

MicroCT images were reconstructed into 3-D models to visualize the Ocn-Cre; Wls^flox/flox skulls from 5 days (Aa and Ab) or 50 days old animals (Ab and Bb). The dorsal views of Col2a1-Cre; Wls^flox/flox embryonic skeletons at E16.5 (Aa and Ab) or E18.5 (Ab and Bb) are shown. The green arrows indicate regions with less mineralization in mutant embryos.

Col2a1-Cre;Wls^flox/flox mutant embryos showed dramatically delayed mineralization at E16.5 and E18.5 (Fig. 2Bb, Bd, Fig. 3 Da, Db). However, Col2a1-Cre;Wls^flox/flox mutant calvarial bone appeared to be normal (Fig. 2 Bb, Fig 3 Da, Db). This is in line with the observation of the absence of Col2a1-Cre activity in calvarial bone (Fig. 1Ba). The mutant Col2a1-Cre;Wls^flox/flox pups had very poorly mineralized ribs, and died shortly after birth. Since Col2a1-Cre;Wls^flox/flox embryos exhibit no changes in viability up to E18.5 (data not shown), we suspect their deaths could be due to pulmonary dysfunction associated with poor mineralization in the ribs. Previous reports utilizing an inactivatable allele of β-catenin (B6.129-Ctnnb1tm2Kem/KnwJ) found that the Col2a1-Cre;β-catenin ^flox/flox and Dermo1-Cre;β-catenin^flox/flox mutant embryos also displayed much poorer mineralization in the ribs and spine compared to the limbs [10, 37].

Dermo1-Cre;Wls^flox/flox mutant embryos did not survive beyond E15.5. They displayed an outgrowth defect in limbs and digit specification (data not shown). Whole mount staining of mutant Dermo1-Cre;Wls^flox/flox embryos showed little mineralization and decreased cartilage content at E14.5 (Fig. 2Cb).

3. Deletion of Wntless inhibits both canonical and noncanonical Wnts to regulate chondrogenesis and osteogenesis in vivo

3.1 Deleting Wls in osteocalcin-positive osteoblasts shows little effect on embryonic bone development

We next investigated the mechanisms underlying the different skeletal phenotypes of Wls mutant embryos. We could detect Ocn-Cre promoter activity in Ocn-Cre;mT/mG embryos as early as E14.5 [Fig. 1 Aa], which is earlier than what was previously reported using an X-gal reporter strategy [27]. However, the lack of noted skeletal patterning changes in mutant Ocn-Cre;Wls^flox/flox embryos could be due to relatively late activation of the Ocn promoter. Considering the fact that the osteoblast-specific Wls deletion by Col1a1-Cre did not show any change in embryonic skeleton either [25, 38], we speculate that osteoblast-derived Wnts may be dispensable for embryonic skeletal development.

3.2 Blocking Wnts from chondrocytes and osteoblasts impairs cartilage development, associated with β-catenin–independent pathway inhibition

In a normal growth plate, chondrocytes in the proliferative zone form ordered columns along the vertical axis of the growth plate. However, H&E staining of mutant Col2a1-Cre;Wls^flox/flox growth plate showed that the columnar cell population in the proliferative zone was disorganized (Fig. 4 B). In addition, we found the same phenotype in mutant spinal discs (Fig. 4 D). These phenotypes are similar to the cartilage phenotypes found in Wnt5a germline knockout embryos [39, 40]. In contrast, we did not find any chondrogenic defect in Col2a1-Cre;Lrp5^flox/flox;Lrp6^flox/flox mutant animals, in which canonical Wnt signaling was selectively inhibited (Joiner, Schumacher and Williams, in preparation). These data are consistent with a model that Wntless deficiency within both cartilage and bone inhibited noncanonical Wnt signaling and impaired cartilage development.

Figure 4. Wntless deletion in Col2 or Dermo1-expressing cells causes defects in chondrogenesis.

The humerus (A and B) or lumbar spine (C and D) from wild-type or mutant E16.5 Col2a1-Cre;Wls^flox/flox embryos were sectioned and stained with H&E or alcian blue. The red arrows indicate the borders of columnar cell population in wild-type group. The humerus (E and F) from wild-type or mutant E13.5 Dermo1-Cre;Wls^flox/flox embryos were also sectioned and stained with H&E. Please note Dermo1 sections (E and F) do not appear to be at a similar level is partially due to the fact that the development of the Dermo1-Cre;Wls^flox/flox embryos is substantially altered. The scale bars indicate 400 μm.

We also observed decreased trabecular mineralization within the mutant Col2a1-Cre;Wls^flox/flox humerus and other bones (Fig. 2Bb, Bd). Since chondrocyte hypertrophy is required for endochondral ossification [41, 42] and Col2a1-cre is also active in osteoblasts and osteocytes (Fig. 1 Ad and Bb), we cannot conclude whether the decreased mineralization in mutant Col2a1-Cre;Wls^flox/flox is caused secondarily by chondrogenic defects or by direct effects of blocking Wnt secretion in osteoblasts/osteocytes.

3.3 Blocking Wnts from chondrocytes and osteoblasts leads to ectopic cartilage, associated with β-catenin-dependent pathway inhibition

Another interesting phenotype of mutant Col2a1-Cre;Wls^flox/flox bone was the presence of ectopic cartilage at the diaphysis of the humerus (Fig. 5 B, arrow). This could either be due either to de novo formation of cartilage or the persistence of a cartilage remnant that failed to be replaced by bone. We observed a similar phenotype in Dermo1-cre;Lrp5 ^flox/flox;Lrp6^flox/flox mutant humeri, but not in Ocn-Cre;Lrp5 ^flox/flox;Lrp6^flox/flox mutant embryos [43]. This ectopic chondrogenesis was also observed in β-catenin conditional knockout embryos with Col2a1-Cre and Dermo1-Cre drivers [37]. We speculate that inhibiting Wnt/β-catenin signaling in osteoprogenitor cells, but not in more differentiated osteoblasts, led to the ectopic cartilage formation. In line with this idea, fewer high-β-catenin-expressing cells were observed in mutant Col2a1-Cre;Wls^flox/flox perichondrium (Fig. 5 D, bracket). In vitro, we found that Wls-deficient and Lrp5/6-doubly-deficient calvarial cells were both more prone to differentiate into chondrocytes and express higher levels of Col2 (Fig. 5 E), and these Wls- or Lrp5/6-doubly-deficient cells also showed less osteogenic potential with lower canonical Wnt signaling in vitro (Fig. 5 E, F) [24, 44].

Figure 5. Wntless deletion in Col2-cre-expressing cells leads to ectopic cartilage and inhibited canonical Wnt signaling.

The humerus from wild-type or mutant E18.5 Col2a1-Cre;Wls^flox/flox embryos were sectioned and stained with alcian blue (A and A, an arrow indicates the ectopic cartilage in the diaphysis) or β-catenin antibody (C and D, yellow bars indicate perichondral region with cells highly expressing β-catenin). Calvarial cells from 3-day-old Wls^flox/flox neonates were infected with adenovirus expressing GFP (Cont) or Cre (Wls−/−) to remove Wls gene. Then cells were differentiated in osteogenic media for the indicated times. Real-time PCR was performed to detect Col2, osteocalcin (E), and Axin2 (F) expression, which was normalized to GAPDH level.

Thus, we conclude that inhibiting canonical Wnt secretion in both chondrocytes and osteoblasts led to an abnormal cartilage formation in Col2a1-Cre;Wls^flox/flox mutant embryos, and that Wnts from both chondrocytes and osteoblasts potentially affect osteoprogenitor cells’ commitment to osteoblastic linage through β-catenin-dependent mechanism.

4. Summary

In this study we evaluated differential requirements of Wnts derived from cells that were descendants of cells that expressed cre from the Osteocalcin promoter (Ocn-Cre), the Col2a1 promoter, or the Dermo1 promoter, and found that both canonical and noncanonical Wnt pathways were involved in the phenotypic changes. However, the fact that we found that Col2a1-cre animals showed evidence for cre-mediated recombination in the mature osteoblasts, means that we could not evaluate the role of Wnts secreted exclusively from chondrocytes.

Our work confirms the work of an earlier manuscript which evaluated skeletal phenotypes of Wls knockout mouse models using a set of similar bone-specific drivers [38]. Although the phenotypes were similar, there were subtle differences noticed in Col2a1-Cre;Wls^flox/flox strain. Firstly, ectopic cartilage formation in Col2a1-Cre driven Wls deletion was not observed [38]. Since we used the same Col2a1-Cre driver, the difference might be due to the timing of observations. We also argue that cell-cell adhesion may not be a major mechanism of ectopic cartilage formation in β-catenin conditional knockout animals, since ectopic cartilage was observed in both Col2a1-Cre;Wls^flox/flox and Dermo1-Cre;Lrp5 ^flox/flox;Lrp6^flox/flox embryos [43]. Secondly, the chondrogenesis defect in proliferative zone was not observed [38]. We initially suspected that might be also due to observations at the different time points. However we found that mutant Col2a1-Cre;Wls^flox/flox humerus also showed a disorganized columnar cell population at E15.5 and E14.5 (data not shown). A more recently published work also showed an ectopic chondrogenesis and an orientation disruption of proliferating chondrocytes in the Col2-driven Wls knockout embryos [45]. So we suspect that the differences seen in ectopic cartilage and columnar cells might be the result of a different mouse genetic background or differences in Col2a1-Cre driver specificities, since the Col2a1-Cre was reported to be chondrocyte-specific in the another study while the Col2a1-Cre we used demonstrated activity in cells that became both chondrocytes and osteoblasts in long bones (Fig. 1Bb).

In summary, we have demonstrated that Wnts from different sources modulate embryonic bone development. Much work is still needed to elucidate the role of Wnts from different tissues, and finding a way to distinguish the roles of canonical or noncanonical Wnts will be useful for dissecting this complex regulation.

Experimental Procedures

Mouse lines

A strain of mice carrying a conditional null allele of Wls^flox (129S-Wls^tm1.1Lan/J) was originally generated by Dr. Richard Lang’s laboratory [24, 46, 47]. Human Ocn-Cre (Tg(BGLAP-cre)1Clem) [27] mice were obtained from the laboratory of Tom Clemens, while Col2a1-Cre(B6;SJL-Tg(Col2a1-cre)1Bhr/J) [28], Dermo1-Cre (B6.129X1-Twist2tm1.1(cre)Dor/J) [29], β-catenin^{flox (}B6.129-Ctnnb1^tm2Kem/KnwJ) and mT/mG (B6.129(Cg)-Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J) transgenic mice were acquired from Jackson Laboratories. The mT/mG reporter mouse possesses loxP sites at both sides of a tdTomato(mT) cassette and expresses red fluorescence in all tissues. When bred to Cre-expressing mice, the resulting offspring have the mT cassette deleted in the Cre-expressing tissues, allowing expression of the enhanced green fluorescent protein (mG) cassette located just downstream of mT. Genomic DNA was prepared from tail biopsies using an AutoGenprep 960 automated DNA isolation system. PCR-based strategies were then used to genotype the mice (details available upon request). All experiments were done in compliance with the guiding principles of the “Care and Use of Animals.” In addition, before use, all procedures were approved by the Institutional Animal Care and Use Committee of the Van Andel Research Institute.

Microcomputed Tomography (microCT) Analysis

Embryonic skeleton was assessed using a desktop SkyScan 1172 microCT imaging system (SkyScan N.V., Vluchtenburgstraat 3C, Belgium). Scans were acquired using a 6.3 μm³ isotropic voxel size, fixed thresholds of 80 and 142, respectively, were used to determine the mineralized bone fraction. Individual CT slices were reconstructed with SkyScan reconstruction software, and used to created 3D models using the software Mimics (Materialise, Leuven, Belgium).

Immunohistochemistry

Bone tissue samples were fixed in 4% paraformaldehyde, and embedded in paraffin. Seven μm sections were adhered to glass slides. Immunohistochemistry was done by Ventana Medical System with antibodies β-catenin (Cell Signaling), GFP (Invitrogen) or Ki-67 (Vector Labs). The whole-mount skeletal staining was done as described previously [48].

Quantitative Real-time RT-PCR Analysis

Total RNA was extracted from cells using RNeasy (Qiagen). The extracted RNA was used for cDNA synthesis via reverse transcriptase using Superscript III and random primers (Invitrogen). The cDNA samples were subjected to PCR analysis using Taqman PCR Master Mix and 20× primer and probes (Applied Biosystems). Amplifications were then performed on an ABI 7500 Real-time PCR system. The expression of the gene of interest and the housekeeping gene (18S ribosomal RNA) were simultaneously determined in the same sample. For each sample, mRNA levels for each gene were normalized to 18s rRNA levels.

Statistical Analyses

Results are expressed as mean ± SD. One-way ANOVA followed by the Bonferroni test was used to measure statistically significant differences between groups. A value of P< 0.05 was considered statistically significant. Data were analyzed using the GraphPad Prism software package (La Jolla, CA).

Abbreviations used

Cre

Cre recombinase

FBS

Fetal Bovine Serum

GFP

green fluorescent protein

GWAS

genome-wide association studies

GAPDH

Glyceraldehyde 3-phosphate dehydrogenase

IHC

immunohistochemistry

LRP

Low-density lipoprotein receptor-related protein

microCT

micro-computed tomography

Ocn

osteocalcin

Wls

Wntless