Specific amelogenin gene splice products have signaling effects on cells in culture and in implants in vivo.

Low molecular mass amelogenin-related polypeptides extracted from mineralized dentin have the ability to affect the differentiation pathway of embryonic muscle fibroblasts in culture and lead to the formation of mineralized matrix in in vivo implants. The objective of the present study was to determine whether the bioactive peptides could have been amelogenin protein degradation products or specific amelogenin gene splice products. Thus, the splice products were prepared, and their activities were determined in vitro and in vivo. A rat incisor tooth odontoblast pulp cDNA library was screened using probes based on the peptide amino acid sequencing data. Two specific cDNAs comprised from amelogenin gene exons 2,3,4,5,6d,7 and 2,3,5,6d, 7 were identified. The corresponding recombinant proteins, designated r[A+4] (8.1 kDa) and r[A-4] (6.9 kDa), were produced. Both peptides enhanced in vitro sulfate incorporation into proteoglycan, the induction of type II collagen, and Sox9 or Cbfa1 mRNA expression. In vivo implant assays demonstrated implant mineralization accompanied by vascularization and the presence of the bone matrix proteins, BSP and BAG-75. We postulate that during tooth development these specific amelogenin gene splice products, [A+4] and [A-4], may have a role in preodontoblast maturation. The [A+4] and [A-4] may thus be tissue-specific epithelial mesenchymal signaling molecules.

Osteogenesis inhibitory protein: a (p)review.

An historical perspective concerning biologically active factors derived from the organic matrix of bone is presented. The emphasis is upon the first discovered of these, bone morphogenetic protein (BMP) and a protein which copurifies with BMP in the initial steps of chemical fractionation. This newly described protein has the biological role of inhibiting the activity of BMP both in vivo and in vitro. Thus it has been named osteogenesis inhibitory protein (OIP). A summary of the known activities of this protein is presented along with preliminary data which indicate the antimitotic nature of OIP.

Isolation, partial purification and in vitro characterization of osteogenic inhibitory protein.

A noncollagenous protein has been extracted and partially purified from adult cortical bone. This protein copurifies with another bone matrix protein, bone morphogenetic protein, until treatment with nonionic detergents. Characterization of the biological activity of this new protein has demonstrated it to be a potent osteogenic inhibitor in vitro. The inhibitor antagonizes the chondrogenic activity of devitalized, demineralized bone matrix as well as the activity of soluble bone morphogenetic protein. Bone matrix induced collagen and glycosaminoglycan synthesis are both inhibited in the presence of various concentrations of the osteogenic inhibitory protein. Inhibition of collagen synthesis required the presence of osteogenic inhibitory protein from the initiation of the tissue culture while glycosaminoglycan synthesis could be inhibited at any stage of differentiation. We postulate that this osteogenic inhibitory protein is essential in normal homeostatic bone metabolism, perhaps acting directly on bone morphogenetic protein.

Purification of bovine bone morphogenetic protein by hydroxyapatite chromatography.

Bovine bone morphogenetic protein (bBMP) induces differentiation of mesenchymal-type cells into cartilage and bone. bBMP has an apparent Mr of 18,500 +/- 500 and represents less than 0.001% of the wet weight of bone tissue. A Mr 34,000 protein resembling osteonectin is separated by extraction with Triton X-100. A Mr 24,000 protein and about half of a Mr 22,000 protein are disassociated from bBMP by precipitation in 1.5 M guanidine hydrochloride. Aggregates of bBMP and a Mr 14,000 protein are insoluble in aqueous media; the bBMP becomes soluble when the Mr 14,000 protein is disassociated in 6 M urea and removed from the solution by ultrafiltration. Three separate molecular species with apparent Mrs 18,500, 17,500, and 17,000 are eluted at 0.10, 0.15, and 0.20 M phosphate ion concentrations, respectively, from a hydroxy-apatite column. The Mr 18,500 protein has the amino acid composition of acidic polylpeptide and includes four half-cystine residues; the pI is 4.9-5.1. The Mr 22,000 component is a chromoprotein resembling ferritin. The NH2-terminal amino acid sequence of the Mr 17,500 protein simulates histone H2B. The Mr 17,000 protein may possess calmodulin activity. Aggregates of the Mr 18,500 and other proteins induce formation of large deposits of bone; the Mr 18,500 protein alone is rapidly absorbed and induces formation of small deposits. None of the other proteins induces bone formation.

Human bone morphogenetic protein (hBMP).

Human bone morphogenetic protein (hBMP) was chemically extracted from demineralized gelatinized cortical bone matrix by means of a CaCl2 X urea inorganic-organic solvent mixture, differential precipitation in guanidine hydrochloride, and preparative gel electrophoresis. hBMP is isolated in quantities of 1 mg/kg of wet weight of fresh bone, and has the amino-acid composition of an acidic polypeptide. The mol wt is 17 to 18 k-Da (kilodaltons). Implants of the isolated 17-kDa protein are very rapidly adsorbed and produce a smaller volume of bone than protein fractions consisting of 24-, 17-, and 14-kDa proteins. Since the isolated 24- and 14-kDA components lack hBMP activity, the kinetics of the bone morphogenetic processes including the function of other proteins as carrier molecules, await investigation.

Basal lamina persistence during epithelial-mesenchymal interactions in murine tooth development in vitro.

Numerous investigations have demonstrated the necessity of mesenchymal instruction for epithelial differentiation during epidermal organogenesis. In the specific case of tooth formation, cap-stage tooth organ mesenchyme instructs epithelial differentiation into ameloblasts with production of enamel extracellular matrix. The "instructive event" is presumed to be direct cell contact. Mesenchyme-mediated cell contact with adjacent epithelia is assumed to "instruct" epithelial differentiation into ameloblasts. If this were true, basal lamina removal and mesenchyme cell contact with epithelia would be prerequisites for epithelial cytodifferentiation and morphogenesis in the developing tooth system. To test this hypothesis, we designed experiments to evaluate basal lamina stability during epithelial differentiation into ameloblasts. Our studies utilized cap-stage murine molar tooth organs, a serumless and chemically defined medium (PYMS), metabolic isotopic labeling of basal lamina constituents, biochemical methods to analyze macromolecular stability throughout 10 days of organ culture in vitro, and immunological methods to localize the distribution of laminin and fibronectin. Our results indicate that (3H)glucosamine is incorporated into basement membranes present in Theiler stage 25 mandibular mouse molar tooth organ. At this stage, the isotope was incorporated into high molecular weight macromolecules. Specific enzyme methods coupled with electrophoresis and fluorography demonstrated that (3H)glucosamine was incorporated into proteoglycans containing chondroitin sulfates, dermatan sulfate, and hyaluronate. After 10 days in vitro the radiolabeled material remained localized in these same molecules, indicating stability of these constituents within basement membranes. Ultrastructural observations indicated that the basal lamina was not removed during ameloblast differentiation in vitro using PYMS medium. Laminin and fibronectin were localized in the basement membranes during cap stages and did not disappear during subsequent morphogenesis and differentiation. Mesenchymal cells appear to mediate epithelial differentiation in vitro using PYMS medium without a removal of the basal lamina.

Fibronectin: its relationship to basement membranes. I. Light microscopic studies.

Fibronectin is a large molecular weight glycoprotein which has been shown to be associated with cell surfaces, extracellular fluids, and connective tissues. Its possible relationship with basement membranes remains controversial. To define this relationship, the distribution of this antigen was evaluated by light microscopic immunoperoxidase techniques in kidney, skin, skeletal muscle, gastrointestinal tract and parietal yolk sac carcinoma. In addition, antibodies against basement membrane and interstitial collagen were used as controls of the specificity of this reaction. Any possible cross-reactivity between plasma fibronectin and basement membrane was examined by immunodiffusion, immunoelectrophoresis and ELISA techniques. The results indicate that antibodies to plasma fibronectin do not co-localize with antibodies to basement membrane constituents. Furthermore, by immunodiffusion or ELISA, there was no cross-reactivity between plasma fibronectin and anti-basement membrane antibody, nor between basement membrane and anti-plasma fibronectin antibody. We conclude that fibronectin is probably not part of the basement membranes studied.

Fibronectin: its relationship to basement membranes. II. Ultrastructural studies in rat kidney.

Fibronectin, basement membrane and type I collagen antigens have been localized in normal rat kidney by electron immunohistochemical methods. Immunoreactive fibronectin was found in the interstitial connective tissue matrix and on collagen fibers, while tubular, endothelial and smooth muscle basement membranes throughout the kidney were consistently negative. In the glomerulus immunoreactive fibronectin was abundant in the mesangial matrix. The peripheral glomerular basement membrane was occasionally reactive in a spotty, irregular manner. These findings suggest that fibronectin antigens are probably not a constituent of basement membranes. It is proposed that some fibronectin antigen may be trapped in the glomerular filter, and that normal glomerular cleansing mechanisms would transport this trapped fibronectin toward the mesangial areas where it would be eventually processed.

Spatiotemporal patterns of fibronectin distribution during embryonic development. I. Chick limbs.

It has been suggested that an extracellular matrix - and cell surface - associated glycoprotein, fibronectin, plays a role in the positioning of cells in morphogenesis and in the maintenance of orderly tissue organization. In the present study the appearance and distribution of fibronectin during in ovo chick limb development has been investigated by indirect immunofluorescence techniques in H.H. stages 20-30. Fibronectin is not detectable until just prior to the transition from the morphogenetic to the cytodifferentiation phase of development. Beginning at H.H. stage 25, successive nonrandom patterns of fibronectin detection and distribution, which resemble the subsequent cartilaginous elements, precede overt chondrogenesis as detected by Alcian blue staining. This corresponds to the onset of the cytodifferentiation phase of limb development. As the accumulation of acidic proteoglycan increases in the cartilage matrix and the mesenchymal cells become more round in appearance, the presence of detectable fibronectin decreases and is ultimately seen only in the perichondria and basement membrane. However, predigestion of developed cartilage tissue with testicular hyaluronidase, prior to fibronectin staining, indicated that fibronectin remains a major constituent of cartilage matrix and is apparently masked by cartilage-specific proteoglycans. This study of chick limb development is consistent with the hypothesis that fibronectin may be a molecule that facilitates the spatial organization of cartilaginous primordia cytodifferentiation.

Possible functions of mesenchyme cell-derived fibronectin during formation of basal lamina.

Epithelial synthesis and secretion of basal lamina has been considered to be a general feature of various vertebrate epithelium-mesenchyme interacting systems (e.g., salivary gland, mammary gland, feather, hair, and tooth morphogenesis). It has been repeatedly assumed that embryonic ectoderm and ectodermal derivatives, such as epithelial tissues associated with tooth morphogenesis, synthesize and secrete basal lamina. Basal lamina of embryonic mouse tooth organs contain laminin, type IV collagen, glycosaminoglycans, and possibly fibronectin. Ectodermally derived epithelia produce laminin, collagens, and glycosaminoglycans but they do not appear to produce fibronectin. Mesenchyme can effect basal lamina formation in vitro by releasing mesenchyme-derived fibronectin. Theiler stage 25 molar tooth mesenchymal and epithelial tissues were enzymatically separated and cultured in chemically defined media without serum, embryonic extracts, or antibiotics for periods not exceeding 24 hr. Isolated epithelia did not reconstitute a basal lamina. Mesenchymepreconditioned media, fibronectin substrata, or addition of 10% fetal calf serum induced reconstitution of epithelium-derived basal lamina. Dental mesenchyme-preconditioned medium contained, as a major component, a protein of M(r) approximately 2.3 x 10(5) identified as fibronectin by the criteria of gelatin binding and subunit molecular weight. Fibronectin was not produced by isolated epithelia. These results support the hypothesis that basal lamina ultrastructural organization results from supramolecular interactions between epithelium-derived macromolecules (e.g., type IV collagen, proteoglycans, glycosaminoglycans, and laminin) with mesenchyme-derived cell surface fibronectin.

Spatiotemporal patterns of fibronectin distribution during embryonic development. II. Chick branchial arches.

It has recently been demonstrated with the chick limb in ovo that successive nonrandom patterns of fluorescent staining with specific antibodies to fibronectin indicate the cartilagenous primordia prior to overt chondrogenesis. Given the apparent nonequivalence of embryonic cartilages, the purpose of this study was to determine whether this phenomenon was unique to developing chick limbs or is a more general characteristic of chondrogenesis. The appearance and distribution of fibronectin during chick first and second branchial arch development in ovo was investigated by indirect immunofluorescence techniques in H.H. stages 15-26. Fibronectin can be detected in early stages in areas presumed to be composed mainly of ectomesenchyme. During later stages of development, successive nonrandom patterns of fibronectin distribution appear to precede overt chondrogenesis as demonstrated by alcian blue staining. Pretreatment of cartilage with testicular hyaluronidase, prior to fibronectin staining, revealed that fibronectin was still present, suggesting that it was masked by proteoglycans. Fibronectin was also detected in the developing membrane bones of the mandible. The nonrandom patterning of fibronectin distribution in ovo in chick branchial arches and limb buds, respectively derived from neural crest and somatic mesoderm, were similar. It appears that specific patterns of fibronectin distribution were characteristic of chondrogenesis, regardless of the embryonic origin of the cartilage. This phenomenon may prove to be an extremely useful probe for early developmental skeletal abnormalities.

DNA synthesis of enamel organ epithelium in vitro is enhanced by co-cultivation with non-viable mesenchyme cells.

Developing enamel organ epithelium was cultured in the presence and absence of non-viable dental papilla mesenchyme cells. Epithelial cell proliferation was enhanced by 75% when cultivated on the mesenchyme cell substratum. The addition of epidermal growth factor (EGF) to enamel organ epithelium cultured on this substratum resulted in a synergistic enhancement of DNA synthesis.

Role of basal lamina in tissue interactions.

Basal lamina are extracellular structures found closely apposed to the plasma membrane on the basal surface of epithelial and endothelial cells and surround muscle and fat tissues. While the primary function of basal lamina in most adult tissues is probably supportive, significant evidence indicates that during embryonic development, basal lamina may be involved in regulating heterotypic tissue interactions. Changes in composition of the basal lamina of salivary and mammary gland epithelial tissues during development probably are important for branching morphogenesis which leads to the final form of the organ. During tooth organogenesis, selective basal lamina degradation and direct cell contact between developing epithelium and mesenchyme has been documented and suggested to be necessary for cytodifferentiation. Basal lamina turnover, as suggested by these morphological observations, no doubt involves both basal lamina synthesis and degradation. We have studied several factors evidently required for basal lamina reconstitution in vitro, and have discovered that fibronectin added to enamel organ epithelial cultures provided cues required for basal lamina formation in vitro. Both fetal calf serum and dental papilla mesenchyme-conditioned media also provided specific signals for basal lamina reconstitution. In addition, we have found that fibronectin is produced by the dental papilla mesenchyme, is released into the medium, and can be isolated from epithelial explants which have been cultured in mesenchyme-conditioned medium; the epithelial explants themselves do not make fibronectin in vitro.

Triple-helix formation on ribosome-bound nascent chains of procollagen: deuterium-hydrogen exchange studies.

Polyribosomes containing nascent [3H]proline-labeled collagen chains were isolated from chick embryo fibroblasts in culture. These nascent chains were nearly completely hydroxylated, as indicated by the presence of [3H]hydroxyproline and high hydroxyproline/proline ratios. The polyribosomes were suspended in D2O at 15 degrees and the infrared spectrum was determined using a reference cell containing collagen-depleted polyribosomes in D2O, matched to equal RNA content. The amide I and amide II bands were observed. When the polyribosomes were heated in D2O at 44 degrees in the infrared cells, the N--D amide II absorbance at 1480 cm-1 increased markedly, indicating that H leads to D exchange had occurred. Collagen-depleted polyribosomes showed no such changes in absorbance at 1450-1480 cm-1 upon heating. Polyribosomes recovered from the infrared cells after treatment at 44 degrees and cooling still contained collagen, as indicated by their [3H]hydroxyproline content. These data indicate that nascent collagen bound to the polyribosomes can assume a hydrogen-bonded structure. Taken with prior data showing that the nascent collagen was also resistant to pepsin digestion, it is suggested that the collagen examined is in triple-helix conformation. Because the nascent polyribosome-bound collagen is nearly fully hydroxylated, it must be considered that triple-helix formation can occur between nascent chains while they are attached to the endoplasmic reticulum surface and that chain association and triple-helix formation in vivo may well occur before rather than after release.

Cell surface carbohydrates of preimplantation embryos as assessed by lectin binding.

Preimplantation embryos were obtained from the uteri and oviducts of 2 strains of mice, Swiss CD-1 and B6CBA. After removal of the zona pellucida by treatment with pronase, FITC-lectins were bound to the embryonic cell surfaces at either 4 degrees C or 37 degrees C. Both morula and blastocyst stage embryos bound the following lectins, FITC-ConA, FITC-WGA, FITC-RCAII and FITC-RCAI. No difference in binding was observed between the morula stage and the blastocyst stage within each mouse strain for each specific lectin. However B6CBA embryos bound less FITC-ConA and FITC-WGA than the corresponding Swiss CD-1 embryos. The topographical arrangement of the lectin receptors was observed to differ between 4 degrees C and 37 degrees C for FITC-ConA, FITC-RCAII, and FITC-RCAI. While lectins bound at 4 degrees C showed a pattern of continuous labeling, the same lectin at 37 degrees C showed aggregation of lectin receptors into patches indicating lateral mobility of these receptors within the embryonic cell membranes. In contrast FITC-WGA bound at 4 degrees C and 37 degrees C demonstrated continuous labeling of embryos at both temperatures. FITC-fucose binding protein did not bind to Swiss CD-1 embryos. The invasiveness of trophoblastic cells of mouse blastocysts was studied by culturing isolated embryos without prior enzyme treatment on reconstituted collagen gels. After 4 days in BME containing only glutamine and bovine serum albumin as supplements, the embryos shed their zona pellucida and implanted into the collagen gel as indicated by zones of lysis in proximity to the embryonic cells when analyzed by scanning electron microscopy.

Intracellular location of triple helix formation of collagen. Enzyme probe studies.

Primary cultures of chick embryo fibroblasts were used to study ribosomal events in the processing of procollagen. Polyribosomes from radiolabeled cells were subjected to enzyme probe analysis using collagenase and pepsin digestion to assess both the amount of procollagen present on the polyribosomes and the conformation of the molecule. The peptides rendered dialyzable by each enzyme treatment were analyzed for radioactive proline and hydroxyproline. Approximately 30% of the nascent proteins were collagenous. Although some hydroxyproline was dialyzable in the pepsin-treated material, a low ratio of hydroxyproline to proline (0.04) indicated that considerable amounts of noncollagenous proteins were digested. Polyribosomal material, previously treated with pepsin, was digested with purified collagenase. Similarly, collagenase-digested polyribosomes were treated with pepsin. The pepsin pretreatment released noncollagenous protein and served to purify the remaining ribosomally bound pepsin-resistant collagenous protein. Collagenase treatment of the pepsin-resistant ribosomally bound peptides released peptides with a hydroxyproline to proline ratio of 0.65, indicating that considerable hydroxylation of proline occurs on nascent ribosomally bound procollagen. This finding combined with the well documented stabilizing effect of hydroxyproline on the collagen triple helix and the demonstrated resistance of ribosomally bound procollagen to pepsin digestion indicates that the collagen triple helix may well form on the polyribosome.