Selective intracellular retention of extracellular matrix proteins and chaperones associated with pseudoachondroplasia.

Mutations in the cartilage oligomeric matrix protein (COMP) gene result in pseudoachondroplasia (PSACH), which is a chondrodysplasia characterized by early-onset osteoarthritis and short stature. COMP is a secreted pentameric glycoprotein that belongs to the thrombospondin family of proteins. We have identified a novel missense mutation which substitutes a glycine for an aspartic acid residue in the thrombospondin (TSP) type 3 calcium-binding domain of COMP in a patient diagnosed with PSACH. Immunohistochemistry and immunoelectron microscopy both show abnormal retention of COMP within characteristically enlarged rER inclusions of PSACH chondrocytes, as well as retention of fibromodulin, decorin and types IX, XI and XII collagen. Aggrecan and types II and VI collagen were not retained intracellularly within the same cells. In addition to selective extracellular matrix components, the chaperones HSP47, protein disulfide isomerase (PDI) and calnexin were localized at elevated levels within the rER vesicles of PSACH chondrocytes, suggesting that they may play a role in the cellular retention of mutant COMP molecules. Whether the aberrant rER inclusions in PSACH chondrocytes are a direct consequence of chaperone-mediated retention of mutant COMP or are otherwise due to selective intracellular protein interactions, which may in turn lead to aggregation within the rER, is unclear. However, our data demonstrate that retention of mutant COMP molecules results in the selective retention of ECM molecules and molecular chaperones, indicating the existence of distinct secretory pathways or ER-sorting mechanisms for matrix molecules, a process mediated by their association with various molecular chaperones.

Fibrillin-1 in human cartilage: developmental expression and formation of special banded fibers.

The molecular basis for Marfan's syndrome (MS), a heritable disorder of connective tissue, is now known to reside in mutations in FBN1, the gene for fibrillin-1. Classic phenotypic manifestations of MS include several skeletal abnormalities associated primarily with overgrowth of long bones. As a first step towards understanding how mutations in FBN1 result in skeletal abnormalities, the developmental expression of fibrillin-1 (Fib-1) in human skeletal tissues is documented using immunohistochemistry and monoclonal antibodies demonstrated here to be specific for Fib-1. At around 10-11 weeks of fetal gestation, Fib-1 is limited in tissue distribution to the loose connective tissue surrounding skeletal muscle and tendon in developing limbs. By 16 weeks, Fib-1 is widely expressed in developing limbs and digits, especially in the perichondrium, but it is apparently absent within cartilage matrix. Fib-1 appears as a loose meshwork of fibers within cartilage matrix by 20 weeks of fetal gestation. Until early adolescence, Fib-1 forms loose bundles of microfibrils within cartilage. However, by late adolescence, broad banded fibers composed of Fib-1 are found accumulated pericellularly within cartilage. Because these fibers can be extracted from cartilage using dissociative conditions, we postulate that they are laterally packed and crosslinked microfibrils. On the basis of these findings, we suggest that the growth-regulating function of Fib-1 may reside persistently within the perichondrium. In addition, the accumulation of special laterally crosslinked Fib-1 microfibrils around chondrocytes during late adolescence suggests that growth-regulating activities may also be performed by Fib-1 at these sites.

Fibrillin-1: organization in microfibrils and structural properties.

To investigate the microfibrillar organization and structural properties of fibrillin-1, we produced overlapping recombinant peptides in human cells which altogether span the fibrillin-1 molecule. The peptides were purified under non-denaturing conditions and extensive characterization indicated correct folding. The purified proteins were used to map monoclonal antibodies 26, 69 and 201. The binding sites are located at the N-terminal end between amino acid residues 45 and 450 (mAb 26), 451 and 909 (mAb 201) and at the C-terminal end between residues 2093 and 2871 (mAb 69). Immunolocalization of these antibodies to extended beaded structures (microfibrils) demonstrated that the N- and C-terminal ends of fibrillin-1 are located in proximity and on opposite sides of the beads, and more central parts of the molecule are located between the beads. Each epitope is present once between each bead. These data allow two possible models for the organization of fibrillin in microfibrils. However, comparison of distances between antibody binding sites on the recombinant peptides and labeling events in tissue suggests that fibrillin molecules are compacted within their tissue form as microfibrils. Additional analysis of the recombinant peptides provide new information regarding the eight-cysteine motif, a novel domain present in fibrillins and TGF beta binding proteins, and suggest that fibrillins are processed at their N-and C-terminal ends.

Elastin exhibits a distinctive temporal and spatial pattern of distribution in the developing chick limb in association with the establishment of the cartilaginous skeleton.

In this work we have analyzed the presence of elastic components in the extracellular matrices of the developing chick leg bud. The distributions of elastin and fibrillin were studied immunohistochemically in whole-mount preparations using confocal laser microscopy. The association of these constituents of the elastic matrix with other components of the extracellular matrix was also studied, using several additional antibodies. Our results reveal the transient presence of an elastin-rich scaffold of extracellular matrix fibrillar material in association with the establishment of the cartilaginous skeleton of the leg bud. The scaffold consisted of elastin-positive fibers extending from the ectodermal surface of the limb to the central cartilage-forming regions and between adjacent cartilages. Fibrillin immunolabeling was negative in this fibrillar scaffold while other components of the extracellular matrix including: tenascin, laminin and collagens type I, type III and type VI; appeared codistributed with elastin in some regions of the scaffold. Progressive changes in the spatial pattern of distribution of the elastin-positive scaffold were detected in explant cultures in which one expects a modification in the mechanical stresses of the tissues related to growth. A scaffold of elastin comparable to that found in vivo was also observed in high-density micromass cultures of isolated limb mesodermal cells. In this case the elastic fibers are observed filling the spaces located between the cartilaginous nodules. The fibers become reoriented and attach to the ectodermal basal surface when an ectodermal fragment is located at the top of the growing micromass. Our results suggest that the formation of the cartilaginous skeleton of the limb involves the segregation of the undifferentiated limb mesenchyme into chondrogenic and elastogenic cell lineages. Further, a role for the elastic fiber scaffold in coordinating the size and the spatial location of the cartilaginous skeletal elements within the limb bud is also suggested from our observations.

Fibrillin binds calcium and is coded by cDNAs that reveal a multidomain structure and alternatively spliced exons at the 5' end.

Fibrillin is an important structural protein of the extracellular matrix. It is a large cysteine-rich glycoprotein with extensive intrachain disulfide bonds, likely contributed by multiple EGF-like repeats. We have previously published 6.9 kb of FBN1 cDNA sequence. FBN1 cDNA clones that extend the sequence 3089 bp in the 5' direction are described in this report. The deduced primary structure suggests that fibrillin is composed of multiple domains. The most predominant feature is the presence of 43 calcium binding EGF-like repeats. We demonstrate here that fibrillin molecules bind calcium. In addition, three alternatively spliced exons at the 5' end are described. Analysis of 5.8 kb of surrounding genomic sequence revealed a 1.8-kb CpG island spanning the alternatively spliced exons and the next downstream exon. Since FBN1 is the gene responsible for Marfan syndrome, the information presented here will be useful in identifying new mutations and in understanding the function of fibrillin in the pathogenesis of the disease.

Four novel FBN1 mutations: significance for mutant transcript level and EGF-like domain calcium binding in the pathogenesis of Marfan syndrome.

Defects of fibrillin (FBN1), a glycoprotein component of the extracellular microfibril, cause Marfan syndrome. This disorder is characterized by marked inter- and intrafamilial variation in phenotypic severity. To understand the molecular basis for this clinical observation, we have screened the fibrillin gene (FBN1) on chromosome 15, including the newly cloned 5' coding sequence, for disease-producing alterations in a panel of patients with a wide range of manifestations and clinical severity. All the missense mutations identified to date, including two novel mutations discussed here, are associated with classic and moderate to severe disease and occur at residues with putative significance for calcium binding to epidermal growth factor (EGF)-like domains. In contrast, two new mutations that create premature signals for termination of translation of mRNA and are associated with reduction in the amount of mutant allele transcript produce a range of phenotypic severity. The patient with the lowest amount of mutant transcript has the mildest disease. These data support a role for altered calcium binding to EGF-like domains in the pathogenesis of Marfan syndrome and suggest a dominant negative mechanism for the pathogenesis of this disorder.

Detection of a progression-linked DNA restriction fragment in rat hepatoma cells probed with a hexokinase cDNA.

We have compared genomic DNA isolated from a series of rat hepatomas and from normal rat liver tissue by restriction fragment analysis. Our objective was to identify a tumor-specific restriction fragment that would be present in rapidly growing tumors and missing in normal tissues. Using a cDNA probe specific for the C-terminal half of rat brain hexokinase, we have identified several changes in DNA restriction fragment lengths in rat hepatomas vs. normal rat liver. Most of these involved restriction fragments that were present in normal rat DNA and missing in tumor DNA. However, one HinfI fragment of M(r) = 13.4 kilobase pairs (kb) was found to be present in rapidly growing rat hepatomas, but missing in normal rat tissue and in hepatomas of slow or intermediate growth rate.

Marfan phenotype variability in a family segregating a missense mutation in the epidermal growth factor-like motif of the fibrillin gene.

To examine the associations among fibrillin gene mutations, protein function, and Marfan syndrome phenotype, we screened for alterations in the fibrillin coding sequence in patients with a range of manifestations and clinical severity. A cysteine to serine substitution at codon 1409 (C1409S) was identified in an epidermal growth factor (EGF)-like motif from one fibrillin allele which segregates with the disease phenotype through three generations of a family affected with the Marfan syndrome. This alteration was not observed in 60 probands from other families or in 88 unrelated normal individuals. The altered cysteine is completely conserved in all EGF-like motifs identified in fibrillin, and in all proteins that contain this motif. These observations strongly indicate that C1409S is the disease-producing mutation in this family. The phenotype of individuals carrying C1409S varied widely with respect to onset of disease, organ-system involvement, and clinical severity; certain affected adults were unaware of their status before being diagnosed through this investigation. We conclude that fibrillin gene defects cause familial Marfan syndrome, that mutations in the EGF-like motif of the fibrillin gene are not uniformly associated with severe disease, and that fibrillin genotype is not the sole determinant of Marfan phenotype.

Partial sequence of a candidate gene for the Marfan syndrome.

Fibrillin is a large (relative molecular mass 350,000) glycoprotein which can be isolated from fibroblast cell cultures and is a component of the microfibrils that are ubiquitous in the connective tissue space. The microfibrils of the suspensory ligament of the lens as well as the elastic fibre microfibrils of the blood vessel wall are composed of fibrillin. The ocular and cardiovascular manifestations of the Marfan syndrome are consistent with a defect in the gene coding for a structural constituent of these connective tissues. Immunohistological experiments have recently implicated fibrillin microfibrils in the pathogenesis of the Marfan syndrome. Genetic linkage data localizing the Marfan gene to chromosome 15 and the in situ hybridization of fibrillin complementary DNA to 15q21.1 together support fibrillin as a candidate Marfan gene. As a first step towards investigating the function of fibrillin in the architecture and development of connective tissues and its relationship to the Marfan syndrome, we report the cloning and partial sequencing of fibrillin cDNA.

Marfan syndrome caused by a recurrent de novo missense mutation in the fibrillin gene.

Marfan syndrome is an inherited disorder of connective tissue manifested in the ocular, skeletal and cardiovascular systems. It is inherited as an autosomal dominant with high penetrance, but has great clinical variability. Linkage studies have mapped the Marfan locus to chromosome 15q15-21.3. There have been no reports of genetic heterogeneity in the syndrome. Following the identification of fibrillin (a glycoprotein component of the extracellular microfibril), immunohistopathological quantification of the protein in skin and fibroblast culture, and examination of fibrillin synthesis, extracellular transport, and incorporation into the extracellular matrix (D. M. Milewicz, R.E.P., E. S. Crawford and P. H. Byers, manuscript in preparation) have demonstrated abnormalities of fibrillin metabolism in most patients. A portion of the complementary DNA encoding fibrillin has been cloned and mapped by in situ hybridization to chromosome 15. Here we report that the fibrillin gene is linked to the Marfan phenotype (theta = 0.00; logarithm of the odds (lod) = 3.9) and describe a de novo missense mutation in the fibrillin gene in two patients with sporadic disease. We thus implicate fibrillin as the protein defective in patients with the Marfan syndrome.

Identification and localization of creatine kinase B and M in normal, ischemic and necrotic myocardium. An immunohistochemical study.

We utilized immunoperoxidase methods to study the distribution of CK-B and CK-M in normal, ischemic and necrotic myocardium. Human myocardium was obtained from autopsy (n = 10) and surgery (n = 16). Cardiac tissue from 22 dogs with experimental myocardial infarction induced by closed-chest coronary balloon occlusion and four dogs with myocardial ischemia without necrosis induced by a 50% reduction in left main coronary artery blood flow for 3 h were studied. Duration of occlusion was 45 min (n = 2), 3 h (n = 8), 5 to 6 h (n = 7), 15 to 24 h (n = 5). Highly purified anti-CK-B and M were prepared in our laboratory and obtained commercially. In all cases, control experiments were performed. Microscopically normal human and dog myocardium uniformly stained for CK-B and CK-M. Necrotic myocardium from patients with acute infarcts (10 to 24 h old) showed markedly reduced immunostaining. In dogs with 3 to 24 h occlusion immunostaining was significantly reduced for both CK-B and CK-M in regions confirmed to be necrotic by triphenyl tetrazolium chloride (TTC) and H & E staining. Myocardial necrosis was confirmed in the 3-h infarcts by electron microscopy (EM). In the four dogs with a 50% reduction in left main flow for 3 h, ischemia was demonstrated by glycogen loss in periodic acid-Schiff stained-sections; but there was no evidence of necrosis by EM or TTC, and there was no loss of immunostaining evident for CK-B and CK-M. Thus, using immunoperoxidase techniques, CK-B and CK-M were visualized in normal and ischemic myocardium, with decreased staining in necrotic tissue. These findings indicate that cell death is necessary for the demonstration of CK-M and CK-B loss from the myocardium by this technique.