|Home | About | Journals | Submit | Contact Us | Français|
Disruption to endochondral ossification leads to delayed and irregular bone formation and can result in a heterogeneous group of genetic disorders known as the chondrodysplasias. One such disorder, multiple epiphyseal dysplasia (MED), is characterized by mild dwarfism and early-onset osteoarthritis and can result from mutations in the gene encoding matrilin-3 (MATN3). To determine the disease mechanisms that underpin the pathophysiology of MED we generated a murine model of epiphyseal dysplasia by knocking-in a matn3 mutation. Mice that are homozygous for the mutation develop a progressive dysplasia and have short-limbed dwarfism that is consistent in severity with the relevant human phenotype. Mutant matrilin-3 is retained within the rough endoplasmic reticulum of chondrocytes and is associated with an unfolded protein response. Eventually, there is reduced proliferation and spatially dysregulated apoptosis of chondrocytes in the cartilage growth plate, which is likely to be the cause of disrupted linear bone growth and the resulting short-limbed dwarfism in the mutant mice.
Chondrocyte proliferation and differentiation within the cartilage growth plate of endochondral bones is the driving force behind longitudinal bone growth. For example, chondrocyte proliferation and matrix deposition in the growth plate is the basis of long bone growth, while apoptosis of hypertrophic chondrocytes plays a pivotal role in the transition from chondrogenesis to osteogenesis (1). The regulation and control of chondrocyte proliferation, hypertrophy and apoptosis in the relevant zones of the growth plate is therefore critical for normal bone growth (2,3) and any disruption to the balance between proliferation and hypertrophy can lead to skeletal defects and in particular the chondrodysplasias (4).
The chondrodysplasias are a clinically and genetically heterogeneous group of diseases that affect the development of the skeleton (5). There are over 200 different phenotypes, which range in severity from relatively mild to severe and lethal forms. Although individually rare, as a group of diseases the chondrodysplasias have an overall incidence of at least 1 per 4000 and result in a significant healthcare responsibility.
Many of the individual phenotypes have been grouped into ‘bone dysplasia families’ on the basis of a similar clinical and radiographic presentation and members of the same family are postulated to share common disease mechanisms (6). Pseudoachondroplasia (PSACH) and multiple epiphyseal dysplasia (MED) are a family of autosomal dominant skeletal dysplasias, which share common phenotypic characteristics but encompass a wide spectrum of severity, ranging from severe to mild phenotypes, respectively (7,8). PSACH results exclusively from mutations in the gene encoding cartilage oligomeric matrix protein (COMP) (9), the fifth member of the thrombospondin protein family. Some forms of MED are allelic with PSACH and also result from COMP mutations, however, MED is genetically heterogeneous and can also result from mutations in the genes encoding matrilin-3 (MATN3) and collagen type IX (COL9A1, COL9A2 and COL9A3) (7). A recessive form of MED is caused by mutations in the sulphate transporter 26A2 protein (SLC26A2/DTDST) (10). Matrilin-3, COMP and type IX collagen are all expressed extensively in a range of skeletal tissues, but in particular throughout the cartilage growth plate during endochondral ossification (11-14).
The matrilins are a family of extracellular matrix (ECM) proteins; matrilin-1 and matrilin-3 are specifically expressed in cartilaginous tissues while matrilin-2 and matrilin-4 have a wider pattern of expression in a variety of ECMs including non-skeletal tissues (15). Matrilin-3 comprises a single von Willebrand factor A-like domain (A-domain), four EGF-like motifs and a coiled-coil oligomerization domain. Matrilin-3 can form hetero-oligomers with matrilin-1 (16,17) and has been shown to bind to COMP and collagen types II and IX in vitro (18,19).
All of the MED mutations identified in MATN3 are missense mutations which primarily affect conserved residues that comprise the β-sheet of the single A-domain of matrilin-3 (20-22); although a single missense mutation has also been identified in the α-1 helix of the A-domain (23). Interestingly, a missense mutation (p.Thr303Met) in the first EGF-domain of matrilin-3 has been implicated in susceptibility to hand osteoarthritis (24,25) and spinal disc degeneration (26) suggesting a role for MATN3 mutations in the pathophysiology of more common cartilage-related diseases (27).
Previous studies of MATN3 mutations using transfected cells as an in vitro assay have suggested that mutant matrilin-3 is retained within the rough endoplasmic reticulum (rER) of chondrocytes (22,28), where it exists as an unfolded intermediate and is associated with ERp72 (22), a chaperone protein known to be involved in mediating disulphide bond formation (29). Although in vivo data is very limited, due primarily to the scarcity of relevant samples, the microscopic analysis of an iliac crest biopsy from a 10-year old boy with MED, caused by a p.Arg121Trp mutation in matrilin-3, shows evidence of enlarged cisternae of rER due to the retention of matrilin-3 (22). In this context, the consequences of MATN3 mutations on the trafficking of matrilin-3 appear similar to those caused by COMP mutations in MED and PSACH (8). For example, mutations in the type III repeats of COMP have been shown to cause the retention of mutant COMP in enlarged cisternae of rER, along with the co-retention of other ECM molecules. This accumulation of protein has been proposed to elicit a cell stress response and results in an apparent increase in apoptosis (30-35).
However, the over reliance on in vitro expression systems to study the effects of MATN3 and COMP mutations has meant that fundamental questions concerning the effect of mutant protein expression on chondrocyte proliferation and apoptosis in the growth plate have remained unresolved. Although a series of knock-out mice have been generated for COMP (36) and the matrilin family of ECM proteins, such as matrilin-3 (37), matrilin-1 (38,39) and matrilin-2 (40), all of these mice have apparently normal skeletal development. More recently, a second matrilin-3-deficient mouse strain was shown to have premature chondrocyte maturation, increased bone mineral density and osteoarthritis, but no chondrodysplasia (41). Overall, these data indicate a functional redundancy within the matrilin and thrombospondin families of proteins and strongly suggest that PSACH and MED are caused by dominant-negative (antimorphic) mechanisms. In order to determine in vivo the disease mechanisms that underlie the pathophysiology of MATN3 mutations we have generated a murine model of MED by introducing a specific human disease-causing mutation (p.Val194Asp) into the A-domain of mouse matrilin-3.
Mice harbouring the equivalent of the human p.Val194Asp mutation in the A-domain of matrilin-3 were generated by homologous recombination in R1 ES cells (Fig. 1A). The GTG → GAT mutation was introduced into exon 2 of matn3 by site-directed mutagenesis. Of 360 ES clones, 33 tested positive for homologous recombination by Southern blotting using the external probe on SpeI digested genomic DNA (Fig. 1B). Twenty-two of which also contained the desired mutation shown by ClaI digestion and direct DNA sequencing (Fig. 1C and D). Six correctly targeted clones were then transiently transfected with the pIC-Cre vector and the deletion of the neo-tk selection cassette in FIAU-resistant ES clones was confirmed by PCR and Southern blot analysis. ES cells from one clone were then used to generate chimeric mice. High contribution chimeric males were mated with C57BL/6 females to generate F1 heterozygote offspring. Further crosses generated mice that were either heterozygous or homozygous for the p.V194D mutation. Normal Mendelian ratios were observed in the offsprings of all matings.
At birth the skeletons of mice heterozygous or homozygous for the p.Val194Asp mutation appeared normal when compared with wild-type littermates (Fig. 2A). At day 21 all mice had similar body weights, however, by day 42 mice that were homozygous for the mutation were approximately 7.5% lighter than either their wild-type or heterozygous littermates and by day 63 these same mice were approximately 9.5% lighter [Fig. 2B; n > 23 mice per genotype, **P < 0.01 by one-way analysis of variance (ANOVA)]. The body weights of mice heterozygous for the mutation were indistinguishable from those of their wild-type littermates.
Bone length measurements were performed on mice of all three genotypes at days 14, 21, 42 and 63. Bone measurements were made of the inner canthal distance (ICD), a measure of intramembranous bone growth; and of the humerus, pelvis, femur and tibia, as measures of endochondral bone growth (Fig. 2C, D and E; only ICD and tibia measurement are shown but the results are representative of all bones measured). At day 14 mice that were homozygous for the mutation had tibia lengths that were approximately 8.5% shorter than both their wild-type and heterozygous littermates (Fig. 2D) and by day 21 the reduction in tibia lengths had progressed to approximately 12.5% (Fig. 2E; n > 6 mice per genotype, **P < 0.01 by one-way ANOVA). No differences were observed in the ICD of age-matched mice of all three genotypes at all ages confirming that intramembranous bone growth was not affected by the matn3 mutation (Fig. 2D and E). These morphometric measurements demonstrated that mice, which are homozygous for p.Val194Asp, are normal at birth but from day 14 develop a measurable short-limbed dwarfism as a result of disturbed endochondral ossification. The age of onset and progressive nature of the dwarfism in mutant mice is comparable to the age of onset in patients with MED, which can be as early as 2 years of age.
Haematoxylin and eosin (H&E) staining of the tibia growth plates from wild-type animals at all ages showed a well organized growth plate in which the resting, proliferative and hypertrophic zones were clearly distinguishable (Fig. 3A). Furthermore, the cells in the proliferative zone were closely aligned in well-ordered columns that were evenly spaced along the horizontal axis of the growth plate. The tibia growth plates from mice homozygous for the matn3 mutation also appeared normal at birth. However, from day 7 mice that were homozygous for the mutation developed a progressively dysplastic growth plate in which the proliferative zone had a disordered cellular organization and morphology (Fig. 3A: inset). This dysplasia was also seen in the hypertrophic zone but no growth plate dysplasia was evident in mice that were heterozygous for the mutation.
Immunohistochemical (IHC) staining of tibia cartilage from newborn wild-type mice showed that matrilin-3 was present throughout the ECM of the growth plate and by day 21 had a predominantly territorial localization. In contrast, analysis of cartilage from newborn mice that were homozygous for the matn3 mutation demonstrated that while matrilin-3 was present in the ECM, there were also detectable levels retained within the pre-hypertrophic and hypertrophic chondrocytes (Fig. 3B). By day 7 the retention of matrilin-3 had become more widespread and was present in chondrocytes from all zones of the growth plate. Finally, by day 21 the retention of mutant matrilin-3 was extensive and the levels of secreted protein were markedly reduced. In contrast, chondrocytes from mice that were heterozygous for the mutation showed no apparent retention of matrilin-3 until day 21 and at this age it was only present in hypertrophic chondrocytes (Fig. 3B). IHC analysis using anti-matrilin-1, -COMP, -collagen II, -collagen IX, -collagen X, -aggrecan, revealed no apparent disruption to the localization of these proteins in the ECM of mice homozygous for the mutation (Supplementary data).
Transmission electron microscopy (TEM) was performed along the entire vertical axis of the growth plate to generate a complete montage of the cartilage growth plate of a 7-day-old mouse tibia (i.e. from resting to terminal hypertrophic and mineralization zones). By aligning images it was possible to compare directly the ultrastructure of a wild-type mouse with littermates that were either heterozygous or homozygous for the mutation (Fig. 4A). In the wild-type growth plate the resting chondrocytes were evenly spaced, while in the early to late proliferative zones, groups of flattened cells (i.e. 2, 4 and 8 cells per chondron) were apparent. The morphology of chondrocytes in the growth plate of mice that were homozygous for the mutation was strikingly different. For example, chondrocytes in the resting zone contained numerous dilated cisternae of rER in addition to the normal rER. These dilated cisternae gradually become larger in size as the chondrocytes underwent proliferation and in some cells they occupied large sections of the cytoplasm, indeed there was a proportion of the proliferating chondrocytes that were ‘wedge-shaped’ in appearance and were unable to align closely within the 4 and 8 cell chondrons. All chondrocytes in the growth plate of mutant mice (n > 200) showed enlarged dilated cisternae of rER, which was not seen in chondrocytes from wild-type mice. Most chondrocytes from mice heterozygous for the mutation had some distended rER, but this was not as prominent as that seen in homozygous mice.
Hypertrophic chondrocytes from wild-type mice were somewhat rectangular in shape and the condensation of chromatin was clearly apparent (42). In contrast, hypertrophic chondrocytes from mice homozygous for p.Val194Asp mutation were more oval in shape and showed the persistence of protein within the remnants of the rER (Fig. 4A).
TEM analysis of the inter-territorial matrix from the proliferating and pre-hypertrophic zones of day 7 tibia growth plates was used to study the ECM architecture of the growth plate cartilage (Fig. 4B). The growth plate cartilage in wild-type mice contained multiple randomly arranged collagen fibrils, but individual fibrils were not always apparent because of a coating of stained proteins (e.g. proteoglycans) (43). In contrast, the collagen fibrils in the ECM of mice homozygous for the mutation were readily identified by their uniform diameter and banding pattern. The differences between collagen fibrils in wild-type and mutant ECM was most apparent when the fibrils were in transverse section (i.e. cut end-on), when the fibrils appeared as dense circles. For example, in the proliferative zone, the mutant samples showed high contrast between the fibril surface and the inter-fibrillar space, whereas in wild-type the abundance of fibril-associated material blurred the fibril boundaries. This effect was most pronounced in the pre-hypertrophic zone; in the mutant samples the fibrils are readily seen whereas in the wild-type, the presence of surface-associated material obscured virtually all of the fibril cross-sections.
Three-dimensional reconstructions of 140 serial TEM sections from the tibia of wild-type and mutant mice at day 7 were used to generate virtual representations of 4-cell chondrons from the proliferative zone of the growth plate (Fig. 5 and supplemental video). In the wild-type mice the volume of the rER, expressed as a proportion of the total cell volume, was 1.3% per individual chondrocytes (average 7.28 μm3/549.25 μm3 cell volume; n = 4). However, in mutant mice the average volume of the distended rER per individual chondrocyte was 11.6% (average 55.3 μm3/476.5 μm3 cell volume; n = 4), thus representing an approximately 9-fold increase in the volume of the rER (Fig. 5 and supplemental video).
The increased expression of the rER chaperones BiP and Grp94 are classical markers of an unfolded protein response activation in both yeast and mammalian cells (44-46). Therefore, to determine if the retention of mutant matrilin-3 was eliciting an unfolded protein response we performed quantitative real-time RT-PCR (qRT-PCR), western blot analysis and IHC to evaluate the relative levels of these chaperone proteins in the chondrocytes of wild-type mice and mice homozygous for the mutation. qRT-PCR analysis confirmed that by 3 days of age the relative levels of BiP and Grp94 were upregulated in mutant chondrocytes by >2- and >4-fold, respectively (Fig. 6A; n = 3 samples per genotype, independent t-test **P < 0.05 and **P < 0.01, respectively), which was confirmed by western blot analysis (Fig. 6B; n = 6 samples per genotype) and IHC (not shown).
When misfolded/unfolded mutant proteins accumulate in the rER they can induce an unfolded protein response, which may cause rER/cell stress and the increased expression of C/EBP homologous protein (CHOP/GADD153) (47). CHOP is a member of the C/EBP family of bZIP transcription factors and because its over expression induces apoptosis we investigated the relative levels of CHOP expression in mutant chondrocytes by qRT-PCR. However, at 3, 5 and 21 days of age there was no detectable increase in the relative levels of CHOP expression compared with wild-type chondrocytes (Fig. 6C).
In order to understand the physiological effect of upregulated chaperone protein expression and the unfolded protein response on chondrocyte differentiation and eventually longitudinal bone growth we determined the relative levels of chondrocyte proliferation in the growth plate. To quantify the levels of chondrocyte proliferation in the growth plate BrdU labelling experiments were performed at day 21. Within the proliferative zone of wild-type mice, 22% of cell nuclei were labelled with BrdU, while only 18% of nuclei were labelled within the proliferative zone of mice homozygous for the mutation, thus signifying an overall decrease of 16% in the rate of chondrocyte proliferation (Fig. 7A; n > 36 sections from >3 mice per genotype, independent t-test **P < 0.01). Proliferation rates in mice heterozygous for the mutation were comparable to those of wild-type mice (data not shown).
To determine if the unfolded protein response was having a detrimental effect on cell viability we determined the relative levels of apoptosis in the growth plate at 21 days of age by counting the number of TUNEL-positive cells when compared with DAPI-stained cells in the hypertrophic zone. In both the wild-type and mutant growth plates, approximately 0.8% of cells were TUNEL-positive which was within normal limits (48) (Fig. 7B; typically two to three TUNEL-positive cells from 285 to 310 DAPI-stained cells per section, n > 20 sections from >3 mice per genotype). However, we noticed that in the mutant growth plate, unlike wild-type mice, that apoptosis was occurring throughout the hypertrophic zone and was not just limited to terminal hypertrophic chondrocytes at the vascular invasion front (Fig. 7C). For example, in the wild-type growth plate apoptosis was mostly restricted to terminal hypertrophic chondrocytes at the vascular invasion front and typically approximately 0.7% of TUNEL-positive cells were specifically located at the vascular invasion front. In contrast, only 0.38% of TUNEL-positive cells were located at the vascular invasion front in mice homozygous for the mutation (Fig. 7D; n > 20 sections from >3 mice per genotype, independent t-test *P < 0.05). Furthermore, the relative number of TUNEL-positive cells throughout the entire hypertrophic zone (i.e. upper → lower hypertrophic zones) was significantly increased in the mutant growth plate, suggesting that apoptosis was spatially dysregulated (Fig. 7E; n > 20 sections from >3 mice per genotype, independent t-test *P < 0.05).
In this study we have generated a knock-in mouse model to determine the disease mechanisms that underpin the pathophysiology of MED caused by a matrilin-3 mutation. We introduced a matn3 mutation (p.Val189Asp) that resulted in a murine chondrodysplasia characterized by mild short-limbed dwarfism. The equivalent human mutation in MATN3 (p.Val194Asp) has previously been shown to cause an autosomal dominant form of MED in which patients are normal at birth, but develop mild short-limbed dwarfism during childhood (20,49). The final heights of MED patients with MATN3 mutations generally range from the third to seventy-fifth percentile and are therefore often within normal limits (21,49,50). It was therefore not surprising that this mouse model of MED did not have a severe short-limb dwarfism; however, by considering the median bone length measurements in wild-type mice as the fiftieth percentile, a 12% reduction in tibia bone length (seen at day 63 in mice homozygous for the mutation) is below the third percentile (i.e. 4 S.D. below the mean). This observation is therefore consistent with the equivalent human phenotype caused by the p.Val194Asp mutation in which the heights of affected adult males were on the seventy-fifth percentile (49), while an affected child in this family is currently growing at just below the third percentile (Dr Geert Mortier, personal communication).
Although the clinical phenotype in the mouse is only evident from day 14, the histological phenotype is in fact observable from birth. This is characterized by the accumulation of mutant matrilin-3 in the rER, which by day 7 leads to a disruption to chondrocyte morphology and columnar organization within the growth plate. Concurrently, there was an upregulation of BiP and Grp94, which are classical markers for UPR activation (44-46) and by day 21 the UPR and growth plate dysplasia is associated with a reduction in chondrocyte proliferation and spatially dysregulated apoptosis. The differences in the timing of the clinical and histological/molecular manifestations of the murine chondrodysplasia is therefore consistent with the natural history of the human MED phenotype and provides a rationale for the progressive short-limbed dwarfism seen in patients (49). Interestingly, in contrast to the human MED phenotype, which is a dominant disease, both copies of the mutant allele were required for the mice to develop a detectable chondrodysplasia. This might be due, in part, to the different physiology of mice or the inbred nature of the strains that were used to generate this mouse model. Similar discrepancies have been observed in other mice models of human skeletal disease such as achondroplasia (51) and mice that harbour a comp p.Thr585Met mutation (Pirog-Garcia et al., manuscript in preparation) and are suggestive of dominance modification (52,53).
Previous studies to elucidate the structural and functional significance of MATN3 mutations in MED have focused on the use of two different in vitro cell culture systems, namely, primary bovine chondrocytes (28) and a mammalian immortalized cell line (22). In both cases matrilin-3 harbouring MED mutations was retained within the rER of cells and in one system remained associated with ERp72 as an unfolded intermediate (22). Furthermore, electron microscopy of cartilage from an MED patient with a MATN3 mutation showed the presence of dilated cisternae of rER; while IHC analysis confirmed that the retained protein was matrilin-3 (22). However, despite these studies several fundamental questions have remained unresolved. For example, although these in vitro studies have suggested that there is a protein trafficking defect of mutant matrilin-3, neither study was able to demonstrate categorically whether mutant matrilin-3 is present within the cartilage ECM, and if a potentially abnormal ECM is able to provide a suitable environment for chondrocyte proliferation and differentiation. IHC analysis of growth plate cartilage from mice homozygous for the mutation demonstrated that while the majority of the mutant protein was retained intracellularly, a smaller proportion was evenly distributed throughout the ECM. Therefore, these data demonstrate for the first time that mutant matrilin-3 can be secreted to some extent from chondrocytes, however, we were unable to determine if this mutant matrilin-3 was present in the ECM as a homo-oligomer (i.e. [matrilin-3]4) or as a hetero-oligomer with matrilin-1 (i.e. [matrilin-1]2 [matrilin-3]2) and it therefore remains a possibility that mutant matrilin-3 monomers can only be secreted when they form hetero-oligomers with normal matrilin-1 monomers.
The morphology of individual chondrocytes, chondrons and indeed the entire growth plate was clearly disrupted in mice that were homozygous for the mutation. The accumulation of protein within dilated cisternae of rER had the effect of altering the shape of some chondrocytes, which were unable to form compact 4- and 8-cell chondrons during cell proliferation. These in turn were unable to align correctly into a columnar arrangement and gave the overall appearance of a disrupted growth plate. It is likely that this misalignment of proliferating cells would have a detrimental affect on linear bone growth. Furthermore, the observed changes in the structure and properties of the ECM are also likely to affect the integrity of the tissue and have a profound affect on cell motility during the proliferation and realignment of chondrocytes at the 2-, 4- and 8-cell stage (54,55). The three-dimensional reconstruction of a 4-cell chondron from both wild-type and mutant mice elegantly illustrates the extent of the accumulation of mutant protein within the distended rER. It is clear from this model that the approximately 9-fold increase in the volume of distended rER is likely to have an adverse physical effect on the chondrocytes, in addition to the UPR. IHC indicated that the accumulation of mutant matrilin-3 was greatest at day 21 (Fig. 3B), but due to technical constraints the three-dimensional reconstructions were only performed on day 7 mice (Fig. 5), therefore the proportion of distended rER would be even greater by day 21.
Prior to this study it was not known whether the retention of mutant matrilin-3 in the rER of chondrocytes would initiate an UPR in vivo. By using a combination of qRT-PCR, western blot analysis and IHC we have been able to establish for the first time that specific chaperone proteins, which are classical markers for UPR activation (44-46), were upregulated in chondrocytes from mutant mice. It is most likely that the activation of the UPR was in direct response to the retention of mutant matrilin-3 and this is normally associated with several downstream consequences, most notable of which is the upregulation of CHOP, a key mediator or rER/cell stress. Interestingly, our observation that mutant matrilin-3 steadily accrues within the rER of chondrocytes suggests that the UPR (including ER-associated degradation) is insufficient to prevent the accumulation of mutant protein. However, we were unable to detect an increase in the relative levels of CHOP expression.
One key goal for generating a murine model of MED was to answer fundamental questions regarding disease mechanisms within the context of the growth plate. Primarily, what is the mechanistic link between the expression of a mutant protein and the decreased linear bone growth and short-limbed dwarfism? The effect of an UPR (and potential rER/cell stress) on chondrocyte proliferation and apoptosis was therefore determined and these data provided novel insight into potential disease mechanisms. We were able to demonstrate for the first time that the relative levels of chondrocyte proliferation were significantly reduced in the growth plate of mice homozygous for the mutation. Interestingly, we did not detect an overall increase in apoptosis; however, apoptosis was spatially dysregulated in the growth plate and significant numbers of TUNEL-positive cells were identified away from the vascular invasion front. These observations suggested that apoptosis was spatially dysregulated rather than increased and are supported by the qRT-PCR data which confirmed that there was no apparent increase in the relative expression of CHOP in mutant chondrocytes. Bearing in mind that chondrocyte proliferation and hypertrophy in the growth plate is vital for long bone growth, any disruption to these processes are likely to severely delay endochondral ossification and eventually lead to short-limb dwarfism. Our data therefore support the hypothesis that reduced chondrocyte proliferation and spatially dysregulated apoptosis is sufficient to cause the progressive short-limb dwarfism that is characteristic of this form of MED.
In summary, therefore we have generated the first relevant murine model of MED and identified mechanistic links between the expression of a mutant gene product and the resulting growth plate dysplasia. Ultimately, this will pave the way for the development of suitable therapeutic approaches for the treatment of diseases within PSACH-MED bone dysplasia family.
A targeting vector was constructed from overlapping lambda clones isolated from a 129/sv library (56). A neo-tk cassette, flanked by loxP sites, was cloned into BamHI/ClaI site of pBluescript II KS (pBSNeotk). The short arm was generated by PCR (SAF 5′-ttctgctcataccgactgg-3′ and SAR 5′-gcaccagctgtaatcatagta-3′) amplification of a 2 kb fragment of mouse genomic DNA, which was digested with Psp OMI/BclI and cloned into the NotI/BamHI sites of the pBSNeotk vector (pBSNeotkSA). The long arm was made from two composite genomic fragments. The 5′ of the long arm, containing the mutation, was generated by PCR amplification of a 1.5 kb fragment from mouse genomic DNA (SalIF 5′-acgcgtcgacttgtttttctgagaggacttcattc-3′ and Mut + XhoIR 5′-cacgggctcgagcagccacctcattcacctggtcctgcggcctcccatctgtatcgataatagctaccttgg-3′). The forward primer had a SalI restriction site (bold), while the reverse contained the XhoI restriction site (bold) found in exon 2, and the mutation which created a ClaI restriction site (italicized). This was digested with SalI and XhoI and cloned into the XhoI site of pBSNeotkSA, disrupting the original XhoI site from the vector and leaving the XhoI site from exon 2 intact. The correct orientation and presence of the mutation was confirmed by digests and sequencing (pBSNeotkSA-SX). Finally, the rest of the long arm, a 7 kb fragment was removed from the genomic clone via XhoI digestion and cloned into the XhoI site of pBSNeotkSA-SX. The orientation was confirmed leaving a continuous 8.5 kb long arm (pBSNeotk-V194D).
The targeting vector DNA (70 μg) was linearized with NotI and used to electroporate 4×107 R1 embryonic stem (ES) cells, grown on feeders and supplemented with leukaemia inhibitory factor (LIF), using a Bio-Rad gene pulser set at 0.8 kV and 3 μF for 0.1 ms. Selection with 500 μg/ml G418 began after 24 h for 5-6 days. DNA from clones were digested with SpeI and analysed by Southern blot hybridization using an external probe (Fig. 1B). In order to remove the floxed neo-tk cassette from homologously recombined clones in vitro, ES cells were electroporated as before but with 50 μg of pIC-Cre, a mammalian expression vector containing the Cre transcript. Cultures were selected with FIAU (0.2 μM final concentration) for 6 days. Clones were picked, DNA isolated and detection of only the 12 kb fragment showed removal of the neo-tk cassette. PCR amplification across the neo-tk site confirmed removal of the cassette and presence of the single remaining LoxP site. Targeted ES cells were microinjected into DBA blastocysts to generate chimeric mice. Chimeric males were bred with C57Bl/6 females. Offspring heterozygous for the mutation were used to generate the matn3 p.Val194Asp mouse strain.
Skeletal preparations of newborn mice were prepared as described previously (57). Growth curves were produced by measuring body weights of littermates at day 21, 42 and 63. Bone length measurements were taken from X-ray radiographs.
Tissue samples were fixed overnight in ice-cold 10% formalin (histological staining and TUNEL) or 95% ethanol/5% acetic acid (IHC and BrdU). Bone samples were then decalcified in 20% EDTA and embedded in paraffin wax. Antibodies used were anti-matrilin-3 and anti-BrdU (Abcam). IHC was carried out on ethanol/acetic acid-fixed samples. Briefly, endogenous peroxidase activity was quenched by a H2O2/methanol wash, followed by hyaluronidase treatment. Samples were blocked with goat serum and BSA in PBS for 1 h, incubated for 1 h with the primary antibody (in PBS/BSA) and then 1 h with the secondary antibody (biotinylated goat anti-rabbit IgG, Dako Cytomation) in PBS/BSA with goat serum. Slides were then incubated with the ABCcomplex/HRP reagent for 30 min, and developed using DAB. Samples were counterstained with methyl green and mounted with VectaMount. BrdU IHC was performed as above, except that the hyaluronidase step was replaced with an antigen retrieval step (incubated in 4 M HCl for 15 min, then neutralized with 0.1 M borate buffer).
Measurements of in vivo apoptosis was carried out using the DeadEnd™ fluorometric TUNEL system (Promega) and visualized with a Zeiss Axiovision microscope. Nuclei where stained with DAPI while apoptotic cells stained with FITC.
Tibia from day 7 mice were fixed overnight in 4% formaldehyde/2.5% gluteraldehyde in 0.1 M sodium cacodylate buffer, followed by three washes in 0.1 M sodium cacodylate buffer. Samples were then incubated (2 h, 4°C) in a secondary fix of 1% osmium tetroxide, followed by three washes in water. En bloc staining of the sample was carried out by incubation (1 h, 4°C) in 0.5% uranyl acetate, followed by water washes. Samples were then dehydrated through an ascending graded acetone series. The acetone was replaced with two changes of propylene oxide, which in turn was replaced with Spurr’s resin. After several changes the resin was polymerized by incubating at 60°C for 48 h. Sections of 70 nm were cut and stained with 0.3% (w/v) lead citrate and images were taken on a FEI Tecnai 12 Biotwin electron microscope and were recorded on 4489 film (Kodak) and scanned using an Imacon Flextight 848 scanner (Precision Camera and Video). Images from EM serial sections were aligned, reconstructed and visualized in IMOD for Linux (58).
For qRT-PCR rib cages from 3- and 5-day-old mice (wt and m/m) were treated with collagenase for 2 h (Type 1A, 2 mg/ml). The costal cartilage was then dissected from the rib cage and the perichondrium layer removed. The cartilage was then further treated with collagenase for 3 h to digest away the collagen matrix and release the chondrocytes. The chondrocytes were passed through a cell strainer (70 μm) and washed twice with PBS. The cell pellet was resuspended in 500 μl Trizol. Total RNA was then isolated according to manufacturer’s instructions (Invitrogen). cDNA was then generated using random hexamer primers (Superscript III, Invitrogen) and qRT-PCR performed using the SYBR® green PCR method for each chaperone. The following primer sequences were used for BiP (For: 5′-ggcaccttcgatgtgtctcttc-3′ and Rev: 5′-tccatga cccgctgatcaa-3′), Grp94 (For: 5′-taagctgtatgtacg ccgcgt-3′ and Rev: 5′-ggagatcatcggaatccacaac-3′) and 18s RNA (For: 5′-gtaaaccgttgaaccccatt-3′ and Rev: 5′-ccatccaatcggtagcg-3′). For western blot analysis, chondrocytes were isolated as above, but aliquots of 2×105 chondrocytes were prepared and resuspended in 5×SDS loading buffer. These aliquots were run on SDS-PAGE then transferred to nitrocellulase membrane for western blot analysis. Ponceau staining was used to confirm equal loading of total protein isolates. Antibodies to key chaperones associated with the unfolded protein response were used at 1:500 dilutions, namely BiP (Santa Cruz) and Grp94 (Santa Cruz).
ANOVA was used to determine differences within and between groups (used for body weight and bone length comparisons), while differences in chondrocyte proliferation, apoptosis and qPCR were analysed using independent t-test. A P-value <0.05 was considered statistically significant.
The video depicts a three-dimensional reconstruction comparing 4-cell chondrons from wild-type and mutant mice. Cells are colour rendered; nuclei are shown in purple, distended rER in blue and primary cilia in red. The video shows the extent of the accumulation of mutant matrilin-3 within the distended rER.
This work was supported by grants from the Wellcome Trust (071161/Z/03/Z to MDB) the National Institute of Health (RO1 AR49547-01 to MDB, RBH, KEK and DJT) and Deutsche Forschungsgemeinschaft (WA1338/2-4 to RW). This work was undertaken in the Wellcome Trust Centre for Cell-Matrix Research and the Histology and Transgenic Mouse core facilities of the Faculty of Life Sciences at the University of Manchester. We would like to thank Dick Heinegard (Lund) for the COMP antibody and Tim Hardingham (Manchester) for the aggrecan antibody.
Conflict of Interest statement. There are no conflicts of interest to declare.