MCDS is an autosomal dominant disorder, with a mild dwarfed phenotype and is caused by mutations in collagen X. The majority of the mutations identified so far are localized almost exclusively within the NC1 domain, which is responsible for trimerization of the collagen X protein. Little is known about the onset of MCDS, but recently, up-regulation of ER stress has been suggested as an important mechanism promoting the MCDS phenotype. Several studies have shown that the mutated collagen X protein is retained within the ER triggering the UPR, which has proved to be the key pathway responsible for the pathogenesis of the MCDS phenotype. In order to study the consequences of the expressing the MCDS-causing COL10A1p.N617K mutation at the molecular level, we selected HeLa cells as an appropriate cell line for the characterisation of the UPR response, by showing that the three branches of the UPR can be activated by ER stress inducing conditions in a similar manner to that seen in vivo in the MCDS growth plate. Importantly we have also shown that HeLa cells can be transduced with the collagen X cDNA constructs and will express, fold and secrete collagen X into the supernatant.Having established the cellular model for MCDS studies we demonstrated for the first time direct evidence for the retention of mutant collagen X within the ER. Moreover, we demonstrated that the mutant collagen X was degraded via a proteasomal pathway. Nevertheless, the level of ER stress induced by expression of mutant collagen X, based on BiP induction at the protein level, was disappointingly low. We therefore directly compared the level of ER stress induced by the COL10A1p.N617K mutation with that of the chondrodysplasias-causing MATN3p.V194D mutation. The ER stress induced by the matrillin mutation was far greater than that caused by the mutant collagen X. We showed that general protein synthesis was reduced in cells expressing either of the mutant proteins, most likely by the mechanism associated with the phosphorylation of eIF2alpha. Moreover, we showed the mutant matrilin-3 protein was also retained specifically in the ER. However, we could find no evidence for either proteasomal or autophagic/lysosomal degradation of mutant matrilin 3.We tested a broad range of ER stress-relieving compounds on cells expressing mutant collagen X and matrilin 3. Carbamazepine, which was previously shown to reduce ER stress in alpha1-antitripsin deficiency, reduced ER stress in cells expressing the mutant collagen X (but not matrilin 3) by way of enhanced proteasomal degradation of the retained protein. This drug should now be tested in vivo against the MCDS mouse to determine its capacity to reduce disease severity.The results presented within this thesis have contributed to the understanding of how cells deal with mutant collagen X and matrilin-3 proteins. We have identified a potential therapeutic compound that may be of use in the treatment of MCDS. Furthermore, the data presented support the concept that generic approaches to relieving ER stress may not be suitable for treating a broad range of diseases. Treatments may need to be tailored not only in a gene-specific manner but also may need to be tailored to address the differing consequences of different mutations in the same gene.

Abstract The role of sequence variations in genes encoding cartilage extracellular matrix (ECM) proteins were studied in osteoarthritis (OA) and multiple epiphyseal dysplasia (MED). The cartilage collagen genes COL2A1, COL9A1, COL9A2, COL9A3, COL11A1, and COL11A2 were screened for sequence variations in 72 Finnish probands and one US family with primary early-onset hip and/or knee OA. Altogether 239 sequence variations were found, of which 16 were not present in the controls. Seven of the unique variations — four in COL11A1, two in COL11A2, and one in COL2A1 — were studied further, because they resulted in the substitution of conserved amino acids or were predicted to affect mRNA splicing. Association analysis was performed by genotyping 6–12 common polymorphisms from each gene in 72 OA patients and 103 controls; no common predisposing alleles were identified. The results, however, suggest that mutations in the minor cartilage collagen genes can be the cause of OA in a subgroup of OA patients. Two MED families with clinical and radiographic features suggestive of a collagen IX mutation were studied. Mutation screening of COL9A1, COL9A2, and COL9A3 yielded negative results. Instead, an R718W mutation in COMP was identified in both families. Clinical and radiographic overlap between patients with collagen IX mutations and patients with COMP mutations points to a common supramolecular complex pathogenesis. Clinical, radiological and molecular analyses of known MED genes were performed on a cohort of 29 consecutive MED patients. The DTDST mutation was identified in four patients (14%), the COMP mutation in three (10%), and the MATN3 mutation in three (10%). Two new distinct phenotypic entities were identified in patients in whom no mutation was found. The findings suggest that mutations in the above mentioned known MED genes are not the major cause of MED and are responsible for less than half of the cases. The existence of additional MED loci is supported by the exclusion of known loci and finding of the specific subgroups among these patients. The results suggest that genetic defects in ECM genes can predispose to OA and cause MED, even though the major genes involved in both disorders remain to be found.

MCDS is an autosomal dominant skeletal dysplasia disorder caused by mutations in collagen X. In most cases, mutations in collagen X result in a misfolded protein which is retained within the ER of hypertrophic chondrocytes, causing increased ER stress. It has previously been demonstrated that increased ER stress causes hypertrophic chondrocytes to de-differentiate in an attempt to avoid the stress. The altered differentiation results in reduced cell hypertrophy and impaired vascular invasion accounting for reduced bone growth. The presence of increased ER stress in hypertrophic chondrocytes is sufficient to cause the MCDS pathology; therefore reducing ER stress may be beneficial in terms of improving the associated pathology. The autophagy enhancing drug carbamazepine (CBZ) has been shown to be capable of reducing ER stress in cells expressing the MCDS-causing p.N617K collagen X mutation. I show in this thesis that CBZ treatment reduced ER stress in HeLa cells transiently expressing a further 3 MCDS-causing collagen X mutations. I have also demonstrated that CBZ treatment induced the degradation of mutant collagen X proteins either through autophagy or proteasomal degradation depending on the nature of the mutation. The drug was tested in vivo using the p.N617K collagen X mouse model of MCDS. In MCDS mice, CBZ reduced the severity of the disease pathology based on histological analyses, restored hypertrophic chondrocyte differentiation toward normal, increased long bone growth rates and decreased the severity of the hip dysplasia. Gene expression analyses on RNA isolated from microdissected hypertrophic chondrocytes revealed that CBZ shifted the pattern of hypertrophic differentiation markers in MCDS mice toward the wild-type pattern, most likely through its stimulation of gene expression associated with intracellular proteolytic pathways. The results presented in this thesis have contributed to the identification of a potential treatment strategy for MCDS- the stimulation of intracellular proteolysis of mutant collagen X. CBZ is FDA approved for the use of epilepsy and bipolar disorder and has a strong safety record in humans. Therefore CBZ could be a potential treatment strategy for MCDS.

京都大学 / 新制・課程博士 / 博士(医科学) / 甲第24531号 / 医科博第145号 / 新制||医科||10(附属図書館) / 京都大学大学院医学研究科医科学専攻 / (主査)教授 斎藤 通紀, 教授 松田 秀一, 教授 遊佐 宏介 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Mutations in the COL10A1 gene cause metaphyseal chondrodysplasia type Schmid (MCDS) by triggering ER stress and unfolded protein response (UPR). MCDS is characterised by a mild short-limb dwarfism accompanied by expansion of the cartilage growth plate hypertrophic zone (HZ) and altered differentiation of hypertrophic chondrocytes (HCs). ATF6 is one of the UPR mediators, which exists in two isoforms, ATF6α and ATF6β. Activation and up-regulation of ATF6α was a prominent biochemical sign of ER stress in a mouse model of MCDS, COL10a1 p.N617K. Although ATF6β is induced and activated in response to ER stress in a similar fashion to ATF6α, the role and significance of ATF6β in the pathology of many ER stress-associated diseases including MCDS is unknown. Here we utilized a combination of in vitro and in vivo approaches to define the precise role of each isoform of ATF6 in MCDS.To investigate the functions of ATF6α and ATF6β in vitro, we developed a MCDS cell model system (expressing either the wild type collagen X or one of the following MCDS-causing mutant forms of the protein: p.N617K, G618V, Y598D, and NC1del10) in which the expression of either ATF6α or ATF6β was efficiently silenced using siRNAs. ATF6α knockdown in HeLa cells expressing different MCDS-causing mutations suppressed the increased expression of UPR-associated genes such as BiP leading to an elevated ER stress, based on increased XBP1 splicing and/or ATF4 protein. In contrast, ATF6β knockdown did not significantly affect the mutant collagen X-induced increased expression of UPR-associated genes. Furthermore, the ER stress levels were significantly reduced in the ATF6β knockdown MCDS mutant cells based on the lower levels of XBP1 splicing and/or ATF4 protein detected. We then crossed the ATF6α/β knockout mice models with COL10a1 p.N617K mouse model of MCDS to investigate the function of ATF6α and ATF6β in vivo. Ablation of ATF6α in MCDS mice further- reduced the endochondral bone growth rate, further expanded the growth plate hypertrophic zone, and disrupted differentiation of HCs. Therefore, ATF6α appeared to play a chondroprotective role in MCDS as its deficiency caused an increase in the severity of the disease. Of particular note, the level of ER stress was further increased in the absence of ATF6α in MCDS, based on enhanced activities of PERK and IRE1 signalling pathways in compensation for the ATF6α loss. Paradoxically, ablation of ATF6β in MCDS mice reduced the intracellular retention of collagen X protein, and alleviated the ER stress as judged by the attenuated activities of PERK and IRE1 signalling pathways. The reduced ER stress resulting from deficiency for ATF6β in MCDS mice restored the expression of collagen X mRNA towards normal and improved the differentiation of HCs, causing a mark decrease in the expansion of HZ. The results presented within this thesis greatly increased our understanding of the function of ATF6α and ATF6β and their interplay in the pathogenesis of MCDS. We demonstrated an indispensable beneficiary role for ATF6α but a detrimental role for its closely related isoform, ATF6β, in pathology of MCDS. We also showed that the role of ATF6β should not be ignored. These findings may be used to develop a potential therapeutic strategy for MCDS through targeting and enhancing ATF6α-dependent and/or attenuating/blocking of ATF6β-dependent signalling pathways.

Chondrodysplasia of Texel sheep is a newly described recessively inherited disorder distinct from other chondrodysplasias described in sheep. Phenotypically normal at birth, affected lambs develop microscopic lesions as early as 9 days of age, and usually demonstrate gross deformities and markedly reduced rates of bone growth by 2 to 3 weeks. Individual bone growth rates are most severely affected in the proximal bones of the forelimbs. Chondrodysplastic lambs typically have short stature, angular limb deformities, a barrel-shaped chest and a wide-based stance. Gross lesions include tracheal narrowing and contortion, enlarged costochondral junctions, and erosion of articular cartilage in major limb joints. Microscopic lesions are confined to hyaline cartilage, and are characterised by degeneration of the interterritorial matrix and dense perichondrocytic rings consisting predominantly of type VI collagen. These lesions are identical in appearance to those in achondrogenesis 1b and diastrophic dysplasia, two diseases caused by defects of the diastrophic dysplasia sulphate transporter (DTDST) in human beings. An investigation to measure the uptake of radiolabelled sulphate by dermal fibroblasts in vitro did not provide evidence of a defect in the DTDST in chondrodysplastic Texel sheep. A linkage disequilibrium study of ovine chromosomes 1, 5, 6, 13 and 22 using microsatellite DNA markers was unable to identify evidence of a mutation causing this form of chondrodysplasia. Capillary electrophoresis of unsaturated chondroitin sulphate disaccharides demonstrated a relative reduction in the ratio of chondroitin 4-sulphate to chondroitin 6-sulphate in affected animals of all ages. This biochemical feature enables the potential determination of the phenotype of newborn lambs prior to the emergence of gross or microscopic lesions. The pathology of the disease, combined with the findings of the genetic, biochemical and in vitro studies, suggest that a mutation may be present in the CHST11 gene. This gene is a good candidate for future studies aimed at discovering the genetic defect in chondrodysplasia of Texel sheep and developing a test to identify heterozygous animals.

Chondrodysplasia of Texel sheep is a newly described recessively inherited disorder distinct from other chondrodysplasias described in sheep. Phenotypically normal at birth, affected lambs develop microscopic lesions as early as 9 days of age, and usually demonstrate gross deformities and markedly reduced rates of bone growth by 2 to 3 weeks. Individual bone growth rates are most severely affected in the proximal bones of the forelimbs. Chondrodysplastic lambs typically have short stature, angular limb deformities, a barrel-shaped chest and a wide-based stance. Gross lesions include tracheal narrowing and contortion, enlarged costochondral junctions, and erosion of articular cartilage in major limb joints. Microscopic lesions are confined to hyaline cartilage, and are characterised by degeneration of the interterritorial matrix and dense perichondrocytic rings consisting predominantly of type VI collagen. These lesions are identical in appearance to those in achondrogenesis 1b and diastrophic dysplasia, two diseases caused by defects of the diastrophic dysplasia sulphate transporter (DTDST) in human beings. An investigation to measure the uptake of radiolabelled sulphate by dermal fibroblasts in vitro did not provide evidence of a defect in the DTDST in chondrodysplastic Texel sheep. A linkage disequilibrium study of ovine chromosomes 1, 5, 6, 13 and 22 using microsatellite DNA markers was unable to identify evidence of a mutation causing this form of chondrodysplasia. Capillary electrophoresis of unsaturated chondroitin sulphate disaccharides demonstrated a relative reduction in the ratio of chondroitin 4-sulphate to chondroitin 6-sulphate in affected animals of all ages. This biochemical feature enables the potential determination of the phenotype of newborn lambs prior to the emergence of gross or microscopic lesions. The pathology of the disease, combined with the findings of the genetic, biochemical and in vitro studies, suggest that a mutation may be present in the CHST11 gene. This gene is a good candidate for future studies aimed at discovering the genetic defect in chondrodysplasia of Texel sheep and developing a test to identify heterozygous animals.

Chondrodysplasia of Texel sheep is a newly described recessively inherited disorder distinct from other chondrodysplasias described in sheep. Phenotypically normal at birth, affected lambs develop microscopic lesions as early as 9 days of age, and usually demonstrate gross deformities and markedly reduced rates of bone growth by 2 to 3 weeks. Individual bone growth rates are most severely affected in the proximal bones of the forelimbs. Chondrodysplastic lambs typically have short stature, angular limb deformities, a barrel-shaped chest and a wide-based stance. Gross lesions include tracheal narrowing and contortion, enlarged costochondral junctions, and erosion of articular cartilage in major limb joints. Microscopic lesions are confined to hyaline cartilage, and are characterised by degeneration of the interterritorial matrix and dense perichondrocytic rings consisting predominantly of type VI collagen. These lesions are identical in appearance to those in achondrogenesis 1b and diastrophic dysplasia, two diseases caused by defects of the diastrophic dysplasia sulphate transporter (DTDST) in human beings. An investigation to measure the uptake of radiolabelled sulphate by dermal fibroblasts in vitro did not provide evidence of a defect in the DTDST in chondrodysplastic Texel sheep. A linkage disequilibrium study of ovine chromosomes 1, 5, 6, 13 and 22 using microsatellite DNA markers was unable to identify evidence of a mutation causing this form of chondrodysplasia. Capillary electrophoresis of unsaturated chondroitin sulphate disaccharides demonstrated a relative reduction in the ratio of chondroitin 4-sulphate to chondroitin 6-sulphate in affected animals of all ages. This biochemical feature enables the potential determination of the phenotype of newborn lambs prior to the emergence of gross or microscopic lesions. The pathology of the disease, combined with the findings of the genetic, biochemical and in vitro studies, suggest that a mutation may be present in the CHST11 gene. This gene is a good candidate for future studies aimed at discovering the genetic defect in chondrodysplasia of Texel sheep and developing a test to identify heterozygous animals.