SEARCH

SEARCH BY CITATION

Keywords:

  • Pkdcc;
  • embryogenesis;
  • embryos;
  • chondrocytes

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

During long bone development, round proliferative chondrocytes (RPCs) differentiate into flat proliferative chondrocytes (FPCs), and then into hypertrophic chondrocytes (HCs). FPCs and HCs support longitudinal bone growth. Here we show that a putative protein kinase gene, Pkdcc (AW548124), is required for longitudinal bone growth. We originally found Pkdcc expressed in the head organizer, but it is also expressed throughout embryogenesis and in various adult tissues. Pkdcc−/− embryos had no head organizer-related defects, but showed various morphological abnormalities at birth, including short limbs, cleft palate, sternal dysraphia, and shortened intestine. In the long bones of the limbs, only the mineralized regions were shortened, and the cartilage length was normal. In the humerus, Pkdcc was strongly expressed in the early FPCs, and FPC and HC formation was delayed in Pkdcc−/− mutants. Together, these data indicate that Pkdcc encodes a protein kinase that is required for the appropriate timing of FPC differentiation. Developmental Dynamics 238:210–222, 2009. © 2008 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Most of the vertebrate skeleton is formed by endochondral ossification, in which bones are formed from cartilage templates generated by the condensation of mesenchymal cells. This well-orchestrated process consists of phases of cell proliferation, maturation, and apoptosis (de Crombrugghe et al.,2001). Initially, the condensed mesenchymal cells differentiate into chondrocytes, which are round and proliferate actively; they are therefore called round proliferative chondrocytes (RPCs). Induction of the transcription factor Sox9 by BMPs (Healy et al.,1999) is required for RPC development (Bi et al.,1999,2001; Akiyama et al.,2002). For example, Sox9 regulates the expression of type II collagen in RPCs (Lefebvre et al.,1997). At the center of the mass of chondrocytes, the RPCs become flat proliferative chondrocytes (FPCs), which proliferate more actively. FPCs form a layer in which they are arranged like stacks of coins that align unidirectionally; this alignment is essential for determining the direction of bone elongation. Subsequently, cells in the center of the FPC layer exit the cell cycle and differentiate into prehypertrophic and hypertrophic chondrocytes (HCs). HCs synthesize type X collagen, undergo mineralization, and attract blood vessels and osteoclasts. HC development is regulated by the transcription factor Cbfa1/Runx2 (Inada et al.,1999; Takeda et al.,2001), and by the interaction of two signaling molecules, Ihh and PTHrP (Lai and Mitchell,2005). Perichondrial cells, which are adjacent to the HCs, become osteoblasts and form the bone collar. Finally, HCs undergo apoptosis and are replaced by bone cells (Kronenberg,2003). Among these multiple steps, the unidirectional alignment of the FPCs and their differentiation into HCs and concomitant increase in cell volume play central roles in the longitudinal growth of long bones. In spite of the importance of FPCs in bone development, the mechanisms underlying their development from RPCs remain unknown.

Here, we show the involvement of Pkdcc (protein kinase domain containing, cytoplasmic [suggested name by MGI: Mouse Genome Informatics], also known as AW548124) in skeletal development. We have been studying mouse embryogenesis. Previously, we found that Single-stranded DNA-binding protein 1 (Ssdp1, also known as Ssbp3) is a co-activator of the transcription factor Lim1, and is essential for the development of one of the head organizer tissues, the prechordal plate (Nishioka et al.,2005). A search for the downstream genes of Ssdp1 led us to identify Pkdcc as a gene strongly expressed in the head organizer tissues. Pkdcc encodes a putative protein kinase. Unexpectedly, the homozygous mouse mutants for Pkdcc had no obvious abnormalities in head organizer activity, but showed various skeletal defects at birth. The most obvious abnormality was short limbs. In the humerus, Pkdcc was strongly expressed in the FPCs, and the formation of FPCs and subsequent skeletal development was delayed in Pkdcc mutants. Therefore, Pkdcc encodes a putative protein kinase that is required for the correctly timed differentiation of FPCs.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Identification of AW548124

Previously, we found that the co-activator protein Ssdp1 is required for prechordal plate development and subsequent head morphogenesis (Nishioka et al.,2005). To search for the genes acting downstream of Ssdp1 in the prechordal plate and/or in head development, we analyzed the gene expression profiles of wild-type and Ssdp1−/− embryos at E8.0 by microarray. Comparison of the gene expression profiles identified 338 genes that were down-regulated in the Ssdp1−/− embryos. Among them, 248 genes, for which the expression patterns were not previously reported, were subjected to in situ hybridization screening. This screening identified AW548124 as a novel gene expressed in the prechordal plate.

AW548124 Encodes a Putative Protein Kinase

In the Mouse Genome Informatics database (http://www.informatics.ax.org/), AW548124 is described as a kinase-like protein. However, the putative AW548124 protein lacks the N-terminal half of the kinase domain, and has been considered to lack kinase activity. Because genes often generate several different transcripts, we hypothesized that AW548124 might also encode one or more transcripts possessing an entire kinase domain. To test this hypothesis, we screened a λgt11 cDNA library of embryonic day (E) 10.5 mouse embryos with the AW548124 cDNA described in the database as a probe. Among the 15 clones obtained, we determined the entire sequences of 6 that contained long cDNAs (Fig. 1A). This analysis revealed the presence of two AW548124 splice variants, type 1 (4 out of 6 sequenced clones) and type 2 (Fig. 1B). The type 1 transcript consisted of 7 exons, and the encoded protein contained protein kinase subdomains I–VII (Fig. 1A, shaded regions, and B). In contrast, the type 2 transcript consisted of 8 exons, and the longest possible protein encoded by this transcript (the circle in Fig. 1A indicates the first codon of the type 2 transcript) lacked protein kinase subdomains I–III. The first intron of the type 2 transcript was in exon 1 of the type 1 transcript (Fig. 1A, open box). The donor-acceptor sequence of this intron was CG-CG, which is divergent from the ordinary sequence, GT-AG (Rogozin et al.,2005). The type 1 transcript included subdomains I and V, which are ATP-binding and catalytic subdomains, respectively, and are essential for the kinase activity (Hanks and Hunter,1995). Therefore, AW548124 type 1 probably encodes a functional protein kinase. Because the AW548124 protein did not contain characteristic localization domains, such as a signal sequence, transmembrane domain, or nuclear localization signal, AW548124 presumably encodes a cytoplasmic protein kinase. Mouse kinome analysis revealed that AW548124 (or SgK493) is a unique protein kinase that belongs to the eukaryotic protein kinase superfamily but cannot be further classified into any known group (Caenepeel et al.,2004). By consulting Mouse Genome Informatics, we renamed this gene as protein kinase domain containing, cytoplasmic (Pkdcc).

thumbnail image

Figure 1. Structure of Pkdcc/AW548124 transcripts. A:Pkdcc/AW548124 cDNA sequence and deduced amino acid sequence. The shaded sequences and roman numerals indicate the position and number of the protein kinase subdomains. The circled sequence indicates the first codon of the type 2 transcript. The boxed sequence is removed in the type 2 transcript as a first intron. The kinase domain was assigned to amino acid residues 120 to 368 (Caenepeel et al.,2004). B: Genomic organization of the two transcripts. The type 1 and type 2 transcripts consist of 7 and 8 exons, respectively. The boxes indicate exons, and black color indicates the position of the longest open reading frame in each transcript.

Download figure to PowerPoint

Expression of Pkdcc

To examine the possible roles of Pkdcc in development, we first observed its expression pattern in mouse embryos by whole-mount in situ hybridization (Fig. 2). To obtain a detectable signal, we used a full-length type 2 cDNA, which detects both type 1 and type 2 transcripts, as a probe. At E6.5, Pkdcc was expressed in the anterior visceral endoderm (AVE) and anterior primitive streak (Fig. 2A–C). At E7.5, Pkdcc was expressed in the anterior definitive endoderm (ADE) and anterior mesoderm (Fig. 2D–F,F'), but not in the notochord (Fig. 2E). At E8.0, Pkdcc was expressed in the ADE and anterior embryonic mesoderm as well as in the prechordal plate (Fig. 2G–I,H',I'). The expression of Pkdcc in all three tissues with head organizer activities, the AVE, ADE, and prechordal plate, suggested its possible involvement in head organizer development. At later stages, however, Pkdcc was expressed rather broadly. At E8.5, its expression was observed widely in anterior tissues (Fig. 2J), and at the midline of the neural plate in the midbrain region (Fig. 2J,K) as well as the lateral margins of the neural plate posterior to the metencephalic region (Fig. 2J,L). Pkdcc was also expressed weakly in the anterior mesenchyme (Fig. 2K–M).

thumbnail image

Figure 2. Whole-mount in situ hybridization analysis of Pkdcc expression in E6.5–E8.5 embryos (A–M). Left side view of embryos at E6.5 (A), E7.5 (D), frontal view of embryos at E7.5 (E) and E8.0 (G), and dorsal view of an embryo at E8.5 (J). B,C,F,H,I,K,L,M: Transverse sections of embryos shown in A, D, G, and J (approximate position of sections are indicated by lines in A,D,G,J). F',H',I': Higher magnifications of F, H, I. N,O: Northern blot analysis of Pkdcc expression in embryos (N) and adult tissues (O). Each lane contained 10 μg of total RNA (N) and 2 μg of poly(A)+ RNA (O). ADE, anterior definitive endoderm; al, allantois; AVE, anterior visceral endoderm; b, brain; ec, ectoderm; ep, epiblast; h, heart; k, kidney; l, lung; li, large intestine; lv, liver; m, mesoderm; me, mesenchyme; nc, notochord; np, neural plate; p, prostate; pp, prechordal plate; PS, primitive streak; s, spleen; sg, salivary, gland; sm, skeletal muscle; st, stomach; t, testis; tm, thymus; tr, thyroid; u, uterus. Scale bars = (J) 200 μm; (A,D–H,I,K-N) 100 μm; (B,C,F',H',I') 50 μm.

Download figure to PowerPoint

We also examined the Pkdcc expression during embryonic development and in adult tissues by Northern blot analysis (Fig. 2N,O). Again, we used a full-length type 2 cDNA, which detects both type 1 and type 2 transcripts, as a probe, because Northern blots using a oligonucleotide probe specific for the type 1 transcript yielded a signal that was too weak to analyze except for large intestine, heart, liver, and salivary gland (data not shown). Therefore, exact distribution of type 1 and type 2 transcripts is currently unknown. Pkdcc appeared as a broad band of around 2.5 kb and was expressed in all the embryonic stages and adult tissues examined. In adults, Pkdcc was expressed strongly in the heart, liver, and testis and weakly in the brain, spleen, lung, and thymus. Such widespread expression suggested that the role of Pkdcc is not restricted to head organizer development.

Generation of Pkdcc-Null Mutant Mice

To elucidate the function of Pkdcc during development, we generated a Pkdcc mutant mouse by homologous recombination in ES cells. In the Pkdcc-targeted allele, we replaced exon 1, which includes the transcriptional and translational start sites, and a part of intron 1 with a PGK-neo-pA cassette (Fig. 3A). Therefore, the resulting mutation destroys both the type 1 and type 2 transcripts. After transforming TT2 ES cells by electroporation and selecting positive clones with G418, we obtained three clones (clone nos. 30, 37, 38) of homologous recombinants. Proper homologous recombination was confirmed by Southern blot analysis of the ES cell DNAs using 5′ and 3′ external probes (Fig. 3B). These clones were injected into 8-cell-stage embryos to generate chimeric mice. All the clones produced male chimeras, which transmitted the mutant allele to their offspring.

thumbnail image

Figure 3. Targeted disruption of the Pkdcc gene. A: Structure of the Pkdcc gene and targeting strategy. The boxes and numbers indicate exons. Protein coding and non-coding regions are indicated by filled and open boxes, respectively. An approximately 2-kb genomic sequence containing the first exon was replaced by a PGK-neomycin-polyA signal cassette (neo) flanked by loxP sites (filled triangles). Positions of the 5′ and 3′ probes are indicated as 5′ and 3′, respectively. Arrows indicate the positions of primers used for genotyping. Arrowheads indicate the positions of primers for RT-PCR. A, AflII; AC, AWcommon primer; AM, AWMT primer; AW, AWWT primer; pBS, pBluescript. B: Southern blot analysis of a targeted ES cell line. Correct targeting of Pkdcc was confirmed by the presence of a 32.4-kb fragment in the AflII digest of genomic DNA using 5′ and 3′ probes. C: RT-PCR showing the absence of Pkdcc transcripts in E9.5 Pkdcc−/− embryos.

Download figure to PowerPoint

Morphological Abnormalities of Pkdcc Homozygous Mutant Mice

Pkdcc heterozygous mutant mice were viable and fertile without obvious abnormalities in their morphology or behavior. The heterozygous mutants were interbred to obtain homozygotes. The Pkdcc−/− embryos showed no obvious abnormality in external morphology at E8.5, E9.5, or E12.5. The absence of Pkdcc transcripts in the homozygous mutants was confirmed by RT-PCR (Fig. 3C). After their birth, however, the Pkdcc−/− mutants showed abnormal respiration and died within a day (Table 1). Because the Pkdcc−/− mutants derived from all three mutant ES clones showed essentially the same abnormalities, we used only the mice derived from ES clone number 37 for further analysis. The Pkdcc−/− neonates were slightly smaller than wild-type neonates (Fig. 4A, A') and weighed 30.3% less than their wild-type littermates (wild-type 1.65 ± 0.36 g [n = 13]; mutant 1.15 ± 0.11 g [n = 17]).

Table 1. Genotypes of Progenies Recovered From Pkdcc+/− Intercrosses
 +/+ (%)+/− (%)−/− (%)Total
  • a

    P0 neonates died shortly after birth.

E8.5–E12.56 (35.3)8 (47.1)3 (17.6)17
E13.5–E14.515 (25.9)29 (50.0)14 (24.1)58
E15.5–E16.514 (31.1)21 (46.7)10 (22.2)45
E18–5-P028 (28.6)54 (55.1)16 (16.3)a98
Postnatal38 (34.5)72 (65.5)0 (0)110
thumbnail image

Figure 4. Gross phenotypes of the Pkdcc−/− mutants. External views (A,A') and skeletal specimens (B–H, B'–H') of wild-type (A–H) and Pkdcc−/− (A'–H') neonates at P0. Stylopod and zeugopod of the forelimb (C,C'), and hindlimb (E,E'). Autopod of the forelimb (D,D') and hindlimb (F,F'). Mineralization in the second phalanges (p2) was strongly reduced (D', F' arrowhead). The tibia of Pkdcc−/− neonates was curved (E' arrow). Ventral view of cervical vertebra (G,G'). The vertebral bodies of c3 to c6 were significantly reduced in the Pkdcc−/− neonates (G', arrow). Right side view of coccygeal vertebrae (H,H'). The ossified regions of co11 to co15 were absent or small in the Pkdcc−/− neonate. Arrowheads show vertebrae posterior to co8 with a mineralized region. c, cervical vertebra; co, coccygeal vertebra; f, femur; fb, fibula; h, humerus; mc, metacarpal; mt, metatarsal; p1-3, phalange 1-3; r, radius; s, scapula; t, tibia; u, ulna. Scale bars: A-C, A'-C',E,E',H,H', 2 mm; D,D',F,F',G,G', 1 mm.

Download figure to PowerPoint

To examine the morphological abnormalities of the Pkdcc−/− neonates in detail, we prepared a series of histological sections and whole-mount skeletal specimens of P0 neonates and found a number of abnormalities, described below. Throughout this study, littermates were used for all the comparative analyses, and consistent results were obtained from independent litters.

Decreased bone mineralization.

Morphologically, the most obvious abnormality was shortened limbs. The long bones of the stylopod—the humerus (forelimb) and femur (hindlimb), and the zeugopod—the radius and ulna (forelimb), and tibia and fibula (hindlimb), were clearly shorter in Pkdcc−/− than in wild-type neonates (n = 9/9) (Figs. 4C, E, C', E', 5A). In these bones, the mineralized regions, which stained red with alizarin red, were significantly shortened (Fig. 5B). The size and morphology of the cartilage, which stained blue with alcian blue, were not significantly affected (Fig. 5C). The tibia of the Pkdcc−/− neonates was curved (n = 9/9) (Fig. 4E′, arrow). In the autopod, the mineralization of the second phalanges (p2) was absent or severely decreased, and that of p1, p3, and the metacarpals or metatarsals was slightly decreased (n = 5/9) (Fig. 4D, D', F, F'). In the spine, the vertebral bodies were reduced at the cervical and coccygeal regions of Pkdcc−/− neonates (Fig. 4G,H, G',H'). Among the cervical vertebrae, the vertebral bodies of the third (c3) to sixth (c6) cervical vertebra were absent or small in the Pkdcc−/− mutants (Fig. 4G, G'). Among the coccygeal vertebrae, vertebral bodies posterior to eleventh coccygeal vertebra (co11) were absent or small in the Pkdcc−/− mutants (Fig. 4H, H', and data not shown). Thus, mineralization was generally reduced in the Pkdcc−/− mutants.

thumbnail image

Figure 5. Reduction of the mineralized region in the long bones of Pkdcc−/− neonates at P0. A: Length of the long bones. B: Length of the mineralized region in the long bones. C: Length of the cartilage region in the long bones. Bars indicate the average and standard errors. Black bars and white bars represent wild-type and mutant embryos, respectively. Asterisks indicate differences that are statistically significant (P < 0.001). The length of the fibula was not measured because these bones were curved in both the control and mutant neonates.

Download figure to PowerPoint

Cleft palate.

Another characteristic abnormality of the Pkdcc mutants was a cleft palate (n = 15/15). The palate consists of an anteriorly localized small primary palate and posteriorly localized large secondary palate. In the Pkdcc−/− mice, a large cleft was present in the midline of the secondary palate (Fig. 6A, A', B, B'). Consistent with this observation, the palatin bones and the maxilla were deformed and displaced laterally (data not shown). Some of the mutants also had a hole in the basisphenoid bone (n = 6/9) (data not shown), although the relationship of this defect to the cleft palate phenotype is currently unknown.

thumbnail image

Figure 6. Other phenotypes of Pkdcc−/− mice. Comparison of wild-type (A–F) and Pkdcc−/− mutants (A'–F'). Ventral view of the palate (A,A') and frontal section of the head (B,B') of neonates. Arrowheads indicate the cleft in the palate. Front view (C,C') and lateral view (D,D') of the atlas (c1) and the axis (c2). The cartilage of the ventral arch (vt) of the atlas fused with the axis (C', arrowheads) in the mutants. The pit in the atlas at the boundary of the dorsal arch and atlas wing was broader in the mutants (D', arrow). Ventral view of the sternum (E,E'). Pkdcc−/− mutants showed dysraphism of the sternum (E' arrowheads), and lacked the posterior tip of the xiphoid processes (E' arrow). Digestive tract of E16.5 embryos (F,F'). aw, atlas wing; c, caecum; d, dens; da, dorsal arch; nc, nasal cavity; ob, olfactory bulb; pp, primary palate; r, rectum; s, stomach; sp, secondary palate; t, tongue; va, ventral arch; vt, ventral tubercle; xp, xiphoid process. Scale bars = (D,D',F,F'), 2 mm; (A–C,A'–C') 1 mm; (E,E') 500 μm.

Download figure to PowerPoint

Other skeletal abnormalities.

In the cervical vertebrae, the cartilage of the ventral arch of the atlas (c1) was fused with that of the axis (c2) in Pkdcc−/− neonates (n = 7/8) (Fig. 6C, C'). In the atlas, a pit at the boundary of the dorsal arch and atlas wing was broadened in the mutants (n = 7/8) (Fig. 6D, D'). In the sternum, dysraphism (n = 2/9) or asymmetric fusion (n = 3/9) was observed in the Pkdcc−/− mutants (Fig. 6E, E'). In addition, dysplasia of the xiphoid processes was observed (n = 9/9). The posterior tip of the xiphoid process was absent (n = 5/9) or separated bilaterally (n = 4/9) in the Pkdcc−/− mutants (Fig. 6E' arrow, and data not shown).

Shortening of the intestine.

Histological examination of the P0 neonates suggested that all the tissues developed normally in terms of cell differentiation. The expansion of alveoli in the Pkdcc−/− neonates was incomplete compared with that of wild-type neonates (data not shown). This could be a secondary effect of insufficient respiration. Dissection of the Pkdcc−/− neonates revealed that most of the internal organs appeared normal, but the lung was not fully expanded and air had accumulated in the stomach and small intestine (data not shown). In addition, the intestine appeared to be shorter than that of control neonates. The shortened intestine was already evident at E16.5, and the lengths of the intestine between the stomach and the caecum (small intestine) and between the caecum and the anus (large intestine) were both clearly shorter in the Pkdcc−/− than in the wild-type embryos (n = 3/3) (Fig. 6F, F').

Delayed Hypertrophy of Chondrocytes in Pkdcc−/− Embryos

The most obvious abnormality of the Pkdcc−/− neonates was the reduced mineralization of the long bones. To address the mechanism of this abnormality, we focused on the humerus and examined its developmental process in histological sections (Fig. 7). In wild-type embryos, the layer of HCs was clearly observed at the center of cartilage at E13.5 (Fig. 7A) and had expanded at E14.5 (Fig. 7B). At E15.5, calcification and bone formation had taken place at the center of the HC layer (Fig. 7C). In contrast, in the Pkdcc−/− embryos, no HC layer was observed at E13.5 (Fig. 7A') or E14.5 (Fig. 7B'). Rather, HCs were first noticed at E15.5 (Fig. 7C') and formed an easily recognizable layer at E16.5, the HCs (Fig. 7D'). These results suggest that chondrocyte hypertrophy is delayed for two days in Pkdcc−/− embryos compared with wild-type embryos.

thumbnail image

Figure 7. Histological analysis of humerus development in Pkdcc−/− embryos. Humerus sections from wild-type (A–D) and Pkdcc−/− (A'–D') embryos stained with hematoxylin and eosin. The HC layer in C' is indicated by a dashed line, because the boundaries between the PCs and HCs were unclear. p, layer of proliferative chondrocytes; h, layer of hypertrophic chondrocytes; b, bone. Scale bars = 200 μm.

Download figure to PowerPoint

To examine the hypertrophic process of chondrocytes at the molecular level, we performed in situ hybridization (Fig. 8). At E13.5, the wild-type humerus widely expressed Col2a1, a marker for proliferative chondrocytes (PCs) (Fig. 8A). After E14.5, the Col2a1 expression disappeared from the central region, and the Col2a1-negative region had expanded by E16.5 (Fig. 8B–D). In Pkdcc−/− embryos, the entire humerus cartilage expressed Col2a1 at E13.5 and E14.5 (Fig. 8A'–B'). At E15.5, the expression of Col2a1 had decreased slightly at the center of the cartilage (Fig. 8C' arrowhead) and was clearly down-regulated in the central region at E16.5 (Fig. 8D'). Col11a2, another marker for PCs (Sugimoto et al.,1998), showed a similar expression pattern (Fig. 8E–G, E'–G'). Col10a1, a marker for HCs, was expressed in the central region of the wild-type cartilage at E13.5 (Fig. 8H), and at E15.5, the Col10a1 expression domain had separated into two domains (Fig. 8I). In the Pkdcc−/− embryos, no expression of Col10a1 was observed through E14.5, and only weak expression was observed at the center of the humerus at E15.5 (Fig. 8I'). Indian hedgehog (Ihh), a marker for pre-hypertrophic chondrocytes (preHCs) (MacLean and Kronenberg,2005), was expressed in the two domains from E14.5 to E16.5 (Fig. 8J–L). In Pkdcc−/− embryos, no expression was observed at E14.5 (Fig. 8J'), only faint expression was observed, at the center of the humerus, at E15.5 (Fig. 8K', K”), and the expression level was increased and the expression domain expanded at E16.5 (Fig. 8L'). These results suggest that in Pkdcc−/− embryos, the timing of the differentiation of PCs into preHCs and HCs is delayed, although the HCs seem to differentiate normally, eventually.

thumbnail image

Figure 8. Molecular analysis of humerus development in Pkdcc−/− embryos. Expression of Col2a1 (A–D, A'–D'), Col11a2 (E–G, E'–G'), Col10a1 (H, I, I'), Ihh (J–L, J'–L'), and Pkdcc (M, M') examined by in situ hybridization of sections of the humerus from wild-type (A–M, M') and Pkdcc−/− (A'–L', K”) embryos. At E15.5, the expression of Col2a1 was only slightly decreased at the center of the mutant cartilage (C', arrowhead). At E15.5, very faint expression of Ihh was observed at the center of the mutant cartilage (K', arrowheads). The position of the bone primordium is outlined with a dashed line (H,I,I', J–L, J'–L'). K” and M' are higher-magnification images of the boxed areas in K' and M, respectively. p, the layer of proliferative chondrocytes; h, the layer of hypertrophic chondrocytes. Scale bars = 50 μm for K”, 200 μm for the others.

Download figure to PowerPoint

Delayed FPC Differentiation in Pkdcc−/− Embryos

To examine the timing with which Pkdcc regulates long bone development, we analyzed the expression of Pkdcc in the E15.5 humerus by in situ hybridization. The hybridization signal was observed in the PCs but not in the HCs (Fig. 8M). PCs change their morphology from round to flat as differentiation proceeds, and Pkdcc expression was especially strong in the FPCs that were adjacent to RPCs (Fig. 8M'). This result suggests that Pkdcc regulates long bone development in the PCs and not in the HCs.

Because the expression pattern of Pkdcc suggested its involvement in the development of FPCs, we re-examined the morphology of the PCs in histological sections (Fig. 9). At E13.5, wild-type embryos had already formed a clear FPC layer (Fig. 9A, yellow line, a2) and this layer became more evident at E15.5 (Fig. 9B, b2) and E16.5 (Fig. 9C, c2). In the Pkdcc−/− embryos, however, no FPCs were observed at E13.5 (Fig. 9A'), although the cell density appeared to be decreased in the central region compared with the terminal region (Fig. 9A', a'1–a'3). At E15.5, although FPCs were present (Fig. 9B', yellow dashed line), they were mixed with RPCs and HCs, and did not form an obvious layer (Fig. 9B', b'2). A clearly visible FPC layer was observed at E16.5 in the Pkdcc−/− embryos (Fig. 9C', c'2). These results suggest that the differentiation of FPCs from RPCs is delayed in Pkdcc−/− embryos.

thumbnail image

Figure 9. Morphological analysis of proliferative chondrocytes with histological sections. Humerus sections of wild-type (A–C) and Pkdcc−/− (A'–C') embryos stained with hematoxylin and eosin. Green, yellow, and white lines indicate the layer of RPCs, FPCs, and HCs, respectively. Dashed lines in B' indicate a mixing of multiple cell types. Scale bar = 100 μm. The insets (a1, a2, etc.) show higher magnification images of the boxed areas. Representative RPCs, FPCs, and HCs are indicated with green, yellow, and white arrowheads, respectively. The inset b'2 shows the mixed presence of these three cell types. Scale bars in insets = 25 μm. D. E: Comparison of the mitotic indexes of RPCs (D) and FPCs (E) in E15.5 embryos. Black and white bars represent wild-type and Pkdcc−/− embryos, respectively. F: Summary of bone development in wild-type and Pkdcc−/− embryos. See Discussion section for details.

Download figure to PowerPoint

Finally, we analyzed cell proliferation by detecting mitotic cells with an anti-phospho Histone H3 antibody in the RPCs and FPCs of E15.5 embryos. For the RPCs, the mitotic indexes of the mutants were comparable to those of their wild-type littermates (Fig. 9D). For the FPCs, the mitotic indexes of the mutants were significantly lower than those of the wild-type littermates (Fig. 9E). Therefore, the proliferation of FPCs, but not RPCs, was reduced in the Pkdcc−/− embryos.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Role of Pkdcc in Bone Development

The most obvious abnormality in Pkdcc−/− neonates was the shortening of the long bones in the limbs. We focused on the humerus, and compared its development between wild-type and Pkdcc−/− embryos. The results are summarized schematically in Figure 9F. Early proliferative chondrocytes are round and differentiate into FPCs in the central cartilage. In wild-type embryos, FPCs and HCs had already formed at E13.5, while in Pkdcc−/− embryos, these cells were not present until E15.5. Thus, the differentiation of RPCs into FPCs was delayed, suggesting that Pkdcc is essential for the timing of the differentiation of FPCs at E13.5. Consistent with this hypothesis, strong Pkdcc expression was observed in the differentiating FPCs. Because FPCs eventually formed in the absence of Pkdcc, Pkdcc function is not essential for cell differentiation itself, but rather is involved in determining the differentiation timing. Because hypertrophic chondrocytes differentiate from FPCs (Crombrugghe et al.,2002), the delayed formation of FPCs may have secondarily caused a delay in HC formation. Because FPCs and HCs support the elongation of long bones (Crombrugghe et al.,2002), their delayed differentiation probably explains why the mineralized regions of the long bones in Pkdcc−/− P0 mice were shorter than in wild type. The reduced proliferation of FPCs in the Pkdcc mutants may have also contributed to the reduction in the mineralized regions. Although we focused our analysis on humerus development, histological examination revealed that the other long bones in the forelimbs and hindlimbs of the Pkdcc−/− mutants showed essentially the same developmental delay (Y. I. and H. S., unpublished observation).

The Pkdcc mutants also showed abnormalities in various other bones, including the vertebrae, digits, sternum, and palate. Because the vertebrae, phalanges, and sternum are formed through endochondral ossification (Chen,1952; Olsen et al.,2000; Montero and Hurle,2007), the reduced ossification in the vertebrae and digits, and the abnormal morphology in the sternum may have also been caused by a delay in chondrocyte development.

Cause of Neonatal Death in Pkdcc−/− Mutant Mice

Pkdcc−/− neonates breathe abnormally, do not suckle, and die within a day after birth. We speculate that the cause of death is insufficient respiration. Although Pkdcc−/− neonates at P0 open their mouth and move their breast and shoulder as if they were breathing, they show cyanosis and eventually die. The lung expansion of Pkdcc−/− neonates is incomplete, and air bubbles accumulate in the stomach and intestine. Because these are representative abnormalities observed in other mutant mice with a cleft palate (Papaioannou,2005), the cause of insufficient respiration in Pkdcc−/− mice is likely to be the cleft palate.

Comparison With Other Mutant Mice

The abnormalities of the Pkdcc−/− mice are different from those seen in mouse mutants of genes essential for bone differentiation, such as Sox9 (Bi et al.,1999), Bmp5 (Kingsley et al.,1992), Pthrp (Lai and Mitchell,2005), and Cbfa1 (Takeda et al.,2001). These mutants fail to undergo mesenchymal condensation, hypertrophy, or osteogenesis. In contrast, cell differentiation itself was not affected in the Pkdcc−/− mutants, although the differentiation timing was delayed. The most evident abnormality in the Pkdcc−/− neonates was a shortening of the long bones of the limbs, and a cleft palate. Similar abnormalities are observed in mouse mutants of δEF1/ZEB1 (Takagi et al.,1998), Wnt5a (Yamaguchi et al.,1999), and Ror2 (DeChiara et al.,2000; Schwabe et al.,2004), and in Gli2−/− and Gli2+/−;Gli3+/− mutants (Mo et al.,1997). δEF1 suppresses BMP target genes (Kondo et al.,2000), Wnt5a and Ror2 are involved in Wnt/planer cell-polarity signaling (Hikasa et al.,2002; Oishi et al.,2003; Qian et al.,2007; Yamamoto et al.,2008), and Gli2 and Gli3 are involved in hedgehog signaling (Ding et al.,1998; Wijgerde et al.,2002; Bai et al.,2004; Riobo and Manning,2007). Therefore, it is tempting to speculate that Pkdcc is a protein kinase involved in the regulation of one or more of these signaling pathways.

Conclusion

Genetic analysis of Pkdcc mutant mice revealed the involvement of Pkdcc in bone development, but its underlying mechanism remains unknown. Molecular and cellular analyses of the Pkdcc kinase should provide insight into the mechanism of bone development.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Screening of Ssdp1 Downstream Genes

Total RNAs were extracted from Ssdp1−/− and wild-type embryos at the early head-fold stage (E7.5) with an RNeasy Micro Kit (Qiagen), and the gene expression profiles were examined by hybridization to the Whole Mouse Genome Oligo microarray (Agilent Technologies). Microarray analysis was performed by Hokkaido System Science Co. Ltd. (Sapporo, Japan). The genes whose expression was decreased more than 33.1% in Ssdp1−/− embryos compared with wild-type embryos were studied by whole-mount in situ hybridization of E8.0 embryos. Digoxigenin-labeled RNA probes were prepared using the RIKEN Fantom 3 cDNAs (Carninci,2005) as templates.

In Situ Hybridization

Whole-mount in situ hybridization was performed as previously described (Wilkinson,1992) using a digoxigenin-labeled Pkdcc/AW548124 RNA probe (GenBank Accession No. BC086639) with E6.5–8.5 embryos. After hybridization, the stained embryos were fixed for 30 min with 4% paraformaldehyde in phosphate buffered saline (PBS), embedded in paraffin, and sectioned at a thickness of 7 μm. Section in situ hybridization was performed as described (Hogan et al.,1994) using digoxigenin-labeled RNA probes on humerus sections from E13.5–16.5 embryos. Forelimbs were fixed overnight with 4% paraformaldehyde in PBS, embedded in paraffin, and sectioned at a thickness of 7 μm. Col2a1,Col10a1, Col11a2, and Ihh cDNAs were kindly provided by Dr. N. Tsumaki (Horiki et al.,2004).

Cloning of Full-Length Pkdcc/AW548124 cDNAs

To obtain the full-length Pkdcc/AW548124 cDNA, an E10.5 mouse embryo λgt11 cDNA library (Sasaki et al.,1999) was screened using a full-length type 2 cDNA (GenBank Accession No. BC086639) as the probe. Hybridization was performed in hybridization buffer consisting of 0.25 M phosphate buffer, pH 7.4, 0.25 M NaCl, 50% deionized formamide, 0.5% SDS, 200 μg/ml salmon sperm DNA, 10% polyethylene glycol (MW8000), and 10 × Denhardt's solution at 42°C for 16 hr. The hybridized membrane was washed with a solution of 0.2 × SSC and 1% SDS at 65°C for 1 hr. The nucleotide sequence of the full-length Pkdcc/AW548124 cDNA encoding the type 1 transcript was deposited into the DDBJ/GenBank/EMBL database with Accession No. AB436780.

Northern Blot Analysis

For embryonic tissues, the total RNA was isolated from mouse embryos with an RNeasy MiniKit (Qiagen), and Northern blots were prepared using standard procedures (Sambrook and Russell,2001). For adult tissues, mouse MTN Blot and MTN Blot II (BD Biosciences) were used. Membranes were hybridized with 32P-labeled full-length type 2 Pkdcc or β-actin cDNA probes using ULTRAhyb (Ambion) following the manufacturer's instructions. Radioactive signals were measured with a BAS2500 bio-imaging analyzer (Fuji Film).

Generation of Pkdcc Mutant Mice

Pkdcc mutant mice (CDB0537K) were generated as follows. We obtained a C57BL/6 mouse BAC genomic clone, RP23-302H7, containing Pkdcc exons, from BACPAC Resources, Children's Hospital Oakland Research Institute. Using the BAC DNA as a template, a 5′ homologous region ranging from 8,359 to 652 bp upstream of the transcription start site of exon 1, and a 3′ homologous region ranging from 1,635 to 5,149 bp downstream of the end of exon 1 were isolated by long PCR using LA-Taq (TaKaRa, Japan) following the manufacturer's instructions. These genomic fragments were cloned to both sides of a PGK-Neo-pA cassette and were used as a targeting vector. TT2 ES (Yagi et al.,1993) cells were electroporated with linearized targeting vectors, followed by positive selection with G418. ES clones were first screened for homologous recombination by long PCR using LA-Taq. The proper homologous recombination was confirmed for the clones that were positive by PCR, by Southern hybridization of the ES genome using 5′ and 3′ external genomic probes, and the absence of random targeting-vector insertion was also confirmed, with an internal probe for the neomycin-resistance gene. The resulting ES clones were injected into 8-cell-stage embryos to produce chimeric mice. The mice were housed in environmentally controlled rooms at the Laboratory Animal Housing Facility of the RIKEN Center for Developmental Biology, under the institutional guidelines for animal and recombinant DNA experiments.

Genotyping

Genotypes were routinely identified by PCR using the mouse genome as a template. The genomes of adults and embryos were extracted from ear punches and yolk sacs, respectively. A 145-bp fragment was amplified for the wild-type allele with primers AWcommon (5′-CAACGTTACAGACAAGGACCACTC-3′) and AWWT (5′-GCGGGGCATTTGGTGTTTGTTAAC-3′), and a 323-bp fragment was amplified for the mutant allele with primers AWMT (5′-TGAATGGAAGGATTGGAGCTACGG-3′) and AWcommon. PCR conditions were 95°C for 1 min, 30 cycles of 95°C for 30 sec, 60°C for 1 min, and 72°C for 1 min, followed by 72°C for 7 min.

RT-PCR

The total RNAs were prepared from E9.5 wild-type and Pkdcc−/− mutant embryos using Trizol reagent (Invitrogen). One microgram of total RNA was used in a reverse transcription reaction with Ready-to-Go You-Prime First-Strand Beads (GE Healthcare), according to the manufacturer's instructions. The resulting cDNA pools were used for PCR amplification of the Pkdcc and β-actin cDNA fragments. The PCR primers used for Pkdcc were 5′-TCTTCTTCACATACCTCCTGCCAC-3′ and 5′-CTTGTAGAGTTCTGCAGGTACTGC-3′. The primers for β-actin were described elsewhere (Nishioka et al.,2008). The PCR conditions were: 95°C for 1 min; 30 cycles of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 1 min; followed by 72°C for 5 min.

Staining of the Skeleton

The cartilage and bone of P0 neonates were stained with Alcian Blue and Alizarin Red, respectively, as previously described (Hogan et al.,1994). Statistical analysis of the bone length was performed with Prism5 statistical software (GraphPad) using an unpaired, two-tailed t-test.

Histology

Embryos and P0 mice were fixed in 4% paraformaldehyde in PBS for 16 hr and in Bouin's solution for 3 days at 4°C, respectively. Following fixation, the specimens were embedded in paraffin, and sectioned at 7 μm. The sections were stained with hematoxylin and eosin according to standard procedures.

Analysis of Cell Proliferation

Paraffin sections were incubated with anti-phospho-Histone H3 (Ser10) (Upstate, No. 06-570, 1:100 dilution), followed by detection with a Vectorstain ABC-PO kit (Vector Laboratories) and diaminobenzidine. The nuclei were lightly counterstained with hematoxylin. For RPCs and FPCs, 460–922 cells and 77–341 cells, respectively, in a randomly selected area were counted for each embryo. Statistical analysis was performed with Prism5 statistical software (GraphPad) using an unpaired, two-tailed t-test.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We are grateful to the Laboratory for Animal Resources and Genetic Engineering for plasmids, the generation of mutant mice, and housing the mice. We thank Dr. N. Tsumaki for cDNA probes, Dr. Y. Hayashizaki for the Fantom3 cDNA clones, Dr. L. J. Maltias for suggestion of the new gene name, Drs A. Sawada, A. Shimono, M. Matsuyama, and S. Yamamoto for technical advice, and members of the Sasaki Lab for discussions. This work was supported by grants from RIKEN to H.S.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES