The regulation of callus formation during fracture repair involves the coordinate expression of growth factors and their receptors. This article describes the temporal and spatial expression of noggin gene, an antagonist to bone morphogenetic protein (BMP), during the fracture repair process. Noggin expression was examined by means of Northern blotting and in situ hybridization and compared with the expression pattern of BMP-4 in a model of fracture repair in adult mice. Expression levels of noggin messenger RNA (mRNA) were enhanced in the early phase of fracture callus formation. The localization of the noggin mRNA was similar to that of BMP-4 mRNA. Distinct noggin mRNA signals were located predominantly in cells lining the periosteum and the cortical endosteum near the fracture site at 2 days after fracture. At 5, 10, and 21 days after fracture, noggin mRNA was detected in the chondrocytes and osteoblasts in the newly formed callus. The pattern of localization was indistinguishable from that of BMP-4. These results suggest that the noggin/BMP-4 balance could be an important factor in the regulation of callus formation during fracture healing.
FRACTURE HEALING is a process that encompasses inflammation, regeneration, and remodeling. The most crucial step in the process is regeneration, which in bone is known as callus formation. A suitable amount of callus formed in the fracture site is required for rapid healing without loss of skeletal mobility. An excess or limited amount of callus sometimes leads to the delay of repair. A number of cellular factors that stimulate callus formation have been reported. Among them, bone morphogenetic proteins (BMPs) are considered to have the central role in callus formation through their bone induction activity.(1–5) Previously, we reported the spatial and transient expression of BMP-4 messenger RNA (mRNA) in the early stages of fracture healing.(6) The localization of BMP-4 correlates with endochondral and membranous ossification sites before osteoid formation. The results combined with the localization of BMP receptors(7, 8) strongly suggest that BMP-4 plays an important role for callus formation. Although BMP-4 mRNA is expressed transiently, the activity of BMP-4 may persist because it is a more stable molecule. Therefore, it is entirely possible that some molecules control the activity of BMP to prevent the generation of an excess amount of callus at the fracture site.
Noggin, a secreted protein of 222 amino acids, was identified as a dorsalizing factor produced by Spemann's organizer during early embryogenesis.(9) In contrast, an Xenopus assay has shown BMPs to act as mesoderm ventralizers and epidermal inducers.(10–17) More recent studies have shown that noggin is an important factor for somite and neural tissue formation(18–22) in addition to mesoderm dorsalizers(23, 24) because it antagonizes the BMPs. Further investigations of the biological function of noggin revealed that this protein acts as an antagonist to BMPs with high affinity for BMPs in vitro.(25, 26) The results of these studies suggest that noggin may be a potent regulator of callus formation during fracture repair. However, negative regulation of BMP activity has not been established in adult tissues.
The present study was undertaken to investigate the relationship between noggin and BMP-4 during the process of fracture repair. To this end, we examined the temporal and spatial expression of noggin mRNA in bone tissue and compared it to that of BMP-4 in mice with fractures by means of Northern blotting and in situ hybridization.
MATERIALS AND METHODS
Experimental model of fracture healing
A total of 75 male BALB/c mice, 8 weeks old and weighing approximately 23 g, were used in this study. The animals were kept in cages with free access to food and water. Under diethyl ether anesthesia, a midlongitudinal incision was made in the back. The back muscles were then retracted to expose the ribs, and a transverse fracture of a rib was produced by cutting it with scissors and the wound was closed with clips. The experimental procedures were undertaken in compliance with the Guidelines for the Care and Use of Animals as described in the American Journal of Physiology.
Tissue preparation and total RNA extraction
Mice were killed at 0, 2, 5, 10, and 21 days after surgery. Fractured ribs with surrounding soft tissues were removed and one of the samples was fixed with 4% paraformaldehyde (PFA) in 0.1 M of phosphate buffer (PB) and used for in situ hybridization analyses. The remaining ribs were homogenized and total RNA was extracted with the Isogen kit (Nippon Gene Co., Tokyo, Japan) according to the manufacturer's instructions. Fixed samples were dehydrated, defatted with chloroform, decalcified with 19% EDTA (pH 7.4), again dehydrated, and then embedded in paraffin wax. Longitudinal sections of 3-μm thickness were cut with a microtome and mounted on 3-(triethoxysilyl)-prorylamin-coated slides (Merck, Darmstadt, Germany) for hematoxylin and eosin staining and in situ hybridization. The slides were stored at 4°C until use.
Northern blot analysis
Thirty micrograms of the total RNA was separated by electrophoresis on a 1.0% agarose-formaldehyde gel and blotted onto Hybond-N+ membrane (Amersham) for Northern blotting.(27) Filters were hybridized with random-primed [32P]-labeled mouse noggin or BMP-4 complementary DNA (cDNA) fragment probes at 68°C in 0.2% bovine serum albumin (BSA), 0.2% polyvinylpyroridone, 0.2% ficoll, 50 mM Tris-HCl (pH 7.5), 10 mM EDTA, 1% NaDodSO4, 0.5 mg/ml heat denatured salmon sperm DNA, and washed three times with 0.1× SSC and 0.5% NaDodSO4 for 1 h at 68°C. Signals were detected by a Mac-Bas1500 radio-image analyzer (Fuji Film, Tokyo, Japan). After hybridization with noggin probe, the membranes were washed and the signals were measured by autoradiography to confirm that no signal was detected. Gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels.
Complementary RNA probes for in situ hybridization
A 0.32-kilobase (kb) fragment of mouse noggin complementary DNA (cDNA) and a 0.72-kb fragment of mouse BMP-4 cDNA were used as templates to synthesize cRNA probes. Details of the mouse BMP-4 cDNA probe used were reported previously.(6, 28) The mouse noggin cDNA was obtained by reverse-transcription polymerase chain reaction (RT-PCR) using mRNA from mouse rib tissue as the template, which were subcloned into pBluescriptII SK(+) plasmid (Stratagene, La Jolla, CA, USA). The primer set used for PCR was as follows: 5′-CAAGGCAAGAAACAGCGCCT-3′(5′ sense) and 5′-CCCCGAGTTCTAGCAGGAAC-3′(3′ antisense; 127-146 and 425-444, respectively). Nucleotide sequences of cDNA were checked to be identical to mouse noggin cDNA (GenBank, U79163).
For in situ hybridization, plasmid containing noggin cDNA fragments was either linearized with XhoI and transcribed with T3 RNA polymerase to generate a 0.32-kb-long antisense cRNA probe or linearized with EcoRI and transcribed with T7 RNA polymerase to generate a sense cRNA probe. Similarly, plasmid transformed with BMP-4 cDNA fragments was either linearized with XbaI and transcribed with T3 RNA polymerase to generate a 0.72-kb-long antisense cRNA probe or linearized with XhoI and transcribed with T7 RNA polymerase to generate a sense cRNA probe.
In situ hybridization
In situ hybridization was carried out as described previously.(29, 30) Digoxigenin (DIG)-labeled single-strand antisense RNA probes of noggin and BMP-4 prepared with a DIG-labeling kit (Boehringer Mannheim Biochemica, Mannheim, Germany) were hybridized with histological sections at 50°C for 16 h. The hybridization signals were then detected with a nucleic acid detection kit (Boehringer Mannheim Biochemica). The control procedures consisted of (1) hybridization with the sense probes and (2) omission of the use of either the antisense RNA probe or the antidigoxigenin antibody. Neither of the control experiments showed any detectable signals.
Change in the expression levels of noggin and BMP-4 mRNA during fracture repair
Thirty micrograms of total RNA extracted from bone tissue on 0, 2, 5, 10, and 21 days after fracture were blotted and hybridized, first with noggin cDNA probe and then with BMP-4 cDNA probe. The changes in the intensities of signals were compared, with GAPDH mRNA signals serving as an internal control (Fig. 1B). The noggin mRNA signals were detectable on day 0. The expression level of noggin was sharply increased on day 2 and gradually decreased through days 5, 10, and 21 (Fig. 1A). BMP-4 mRNA also was detectable on day 0 and showed an increase on day 2, with slightly decreased expression levels thereafter (Figs. 1A and 1B).
Identification of the cell types expressing noggin and BMP-4
Three serial paraffin sections of 3-μm thickness were prepared from fractured bone tissues. The first was hybridized with noggin cRNA probe; the second was hybridized with BMP-4 cRNA probe; the third was stained with hematoxylin and eosin. Faint and disperse signals of noggin and BMP-4 mRNA were detected at day 0 with no specific localization (data not shown). On day 2 after fracture, proliferation of mesenchymal cells was recognizable in the periosteum and endosteum layers (Fig. 2A). Distinct noggin mRNA signals were localized in the cells of the thickened periosteum layer (Figs. 2B-2D), muscle layer (Figs. 2B and 2C), and endosteum close to the fracture gap (Figs. 2B, 2C, and 2E). BMP-4 mRNA also was detected in the periosteum (Figs. 2F-2H), muscle (Figs. 2F and 2G), and endosteum (Figs. 2F, 2G, and 2I). On day 5, cartilage and woven bone were visible in the periosteal layer (Fig. 3A). Noggin mRNA was detected in the chondrocytes in newly formed cartilage (Figs. 3B, 3C, and 3E) and osteoblasts in the surface of woven bone (Figs. 3B-3D). Localization of BMP-4 mRNA was essentially similar to that of noggin. It was localized to chondrocytes (Figs. 3G, 3H, and 3J) and osteoblasts (Figs. 3G-3I). On day 10, endochondral ossification was clearly visible (Fig. 4A). Noggin and BMP-4 mRNA were colocalized in the chondrocytes and hypertrophic chondrocytes of the fracture gap (Figs. 3B, 3C, 3D, 3G, 3H, and 3I), osteoblasts lining woven bone (Figs. 3B, 3C, 3E, 3G, 3H, and 3J), and mesenchymal cells in the bone marrow (Figs. 4B and 4G). On day 21, the fractured gap was bridged by newly formed bone by endochondral and membranous ossification (Fig. 5A). Noggin and BMP-4 mRNA were colocalized in chondrocytes and hypertrophic chondrocytes (Figs. 5B, 5C, 5D, 5F, 5G, and 5H) and osteoblasts on the trabecular surface in the callus (Figs. 5B, 5C, 5E, 5F, 5G, and 5I). Because the sense-strand probes for noggin and BMP-4 showed no signals (data not shown except Figs. 3F and 3K and 4F and 4K), the signals were judged specific for the transcripts of these mRNAs.
The striking phenotype of excess bone and cartilage formation in the knockout mice carrying a disrupted noggin gene indicates that noggin is one of the key factors regulating skeletal development. In the absence of noggin, BMP is free to exert its bone inducing potential in an unrestricted manner.(31) Taken together with the results of biochemical investigations, noggin clearly shows antagonistic activities toward BMP-2, -4, and -7 homodimers and heterodimers, and must therefore be viewed as a negative regulator of BMPs in vivo. Furthermore, seven mutations have been detected in patients with proximal symphalangism and multiple synostoses syndrome, which manifests as impaired malformation in the joints and vertebra.(32) Therefore, it is thought that the noggin-BMP interaction has an important role in chondrogenesis during development. In spite of these remarkable findings, little information is available about the cell types expressing noggin gene in the skeletal tissues such as bone and cartilage in adult tissues.
The spatial and transient expression of noggin mRNA in the process of fracture healing was investigated in the present study. The up-regulation of noggin gene expression was established by Northern hybridization. After showing a peak on day 2, it gradually decreased to day 21. The change in the expression levels of noggin mRNA was essentially similar to that of BMP-4, which was examined in the present study using a more quantitative procedure than described in our previous report.(6) The localization determined by in situ hybridization was compared with that of BMP-4 mRNA. Careful examination of adjacent sections could not distinguish any difference between the localization of these two mRNAs. They were expressed in cells located in the proliferating periosteal layer, cells lining the newly formed bone (presumably osteoblasts), cartilage tissue including differentiating chondrocytes, and hypertrophic chondrocytes. To the best of our knowledge, this is the first report to analyze the expression of noggin in adult skeletal tissues. The expression of noggin in chondrocytes during the process of fracture healing is almost consistent with previous observations during development.(31, 33–35) Although noggin expression in hypertrophic chondrocytes found in this study was not observed during embryogenesis, recent studies suggest that noggin coordinates chondrogenesis with BMPs.(36, 37) However, no significant expression of the noggin gene has been reported in osteoblastic cells in vivo. The amplified noggin mRNA signal in osteoblastic cells found in this study may be a result of enhanced expression of BMPs and BMP receptors. Previous in vitro studies indicated that BMP-2, -4, and -6 and transforming growth factor β1 (TGF-β1) induced noggin expression in osteoblastic cells.(37) Abundant expression of BMP-4 shown in the previous studies(6, 8, 38) and in this study, together with the colocalization of BMP receptors in the process of fracture healing,(7, 8) support our hypothesis. The reason for the discrepancy between embryogenesis and fracture healing with respect to the expression of noggin in osteoblastic cells is unknown. Conceptually, it is possible to understand the importance of noggin in regulating mineralization mechanisms involved in adult skeletal tissues, pathological calcification such as atherosclerosis, or renal stone formation. Indeed, numerous reports showing the involvement of BMPs in these tissues has been made.(39–43) Understanding the balance between BMPs and noggin in the process of pathological calcification could provide further answers to support this hypothesis. The biological significance of noggin in the homeostasis of adult tissue needs to be elucidated in future studies.
The colocalization of noggin, BMP-4, and BMP receptors in the process of fracture healing found in this study may be quite important in considering the regulatory mechanism of callus formation in fracture repair. It appears that noggin-BMP interactions in adult bone are regulated as an autocrine mechanism. Further investigations of the signal transduction and regulatory pathway involving BMP receptors and noggin gene expression will be required to understand fully this mechanism.
We wish to thank Ms. Kazuko Misawa and Noriko Kiuchi for preparing the histological sections and for their support in this study.