The TUNEL (transferase-mediated, dUTP-biotin nick end labeling) method for in situ labeling of DNA strands was utilized to localize DNA fragmentation in cells involved in tooth formation in the neonatal mouse and hamster. Positive reactions for the presence of DNA fragments were obtained in some epithelial cells of the cervical loop region of incisors, late secretory, transitional and early maturation stage ameloblasts, stratum intermedium cells and in shortened ameloblasts just before eruption. Also, cells of the periodontal ligament of the continuously erupting incisors stained positive shortly before eruption. Odontoblasts were negative but became strongly positive during the formation of physiological osteodentin at the tip of developing incisors. Osteodentin matrix and the surfaces of unerupted enamel and cementum just prior to eruption stained for DNA fragments as well. DNA fragmentation could be elicited in odontoblasts and underlying pulpal tissues of mature erupted molars after mechanical injury to the odontoblast processes during cavity preparation. We conclude that, in rodents, DNA fragmentation and cell death are biological processes which take place in a variety of cells involved in formation of teeth. The TUNEL staining technique is a simple but powerful tool to examine the fate of cells and tissues undergoing either programmed cell death (apoptosis) or fragmentation of nuclear DNA induced by external factors leading to pathological changes.

To study the fate of bone cells, we used the transferase-mediated, biotin-dUTP nick end-labeling (TUNEL) assay to detect DNA fragmentation during the formation of intramembranous and endochondral bone in newly born hamsters, mice, and rats. In alveolar bone forming around the developing tooth crowns, DNA fragmentation was found in three cell types: TRAP-negative mononuclear cells at the bone surface, osteocytes, and some but not all nuclei of TRAP-positive osteoclasts. Osteoblasts did not undergo DNA fragmentation. A strong positive correlation was found between contacts of TUNEL-positive osteocytes and osteoclasts. Extracellular bone matrix also stained occasionally for the presence of DNA fragments. During endochondral bone formation, TUNEL staining was detected in late hypertrophic chondrocytes of the epiphyseal growth plate. During rapid longitudinal growth of long bones, TUNEL-positive hypertrophic chondrocytes were found coincident with or slightly after invasion of blood vessels from the diaphysis. However, during slow longitudinal growth and in secondary ossification centers, DNA fragmentation was seen in hypertrophic chondrocytes still located within their lacunae. We conclude that some of the osteocytes in deeper layers of bone die within their lacuna and disperse nuclear fragments over the extracellular matrix, that a majority of the osteocytes are phagocytosed and degraded by osteoclasts at sites of intense bone resorption, and that during endochondral ossification, substantial numbers of late hypertrophic chondrocyte cells undergo cell death.

Endochondral ossification in the epiphyseal growth plate of long bones is associated with programmed cell death (PCD) of a major portion of the chondrocytes. Here we tested the hypothesis that at the ossification front of the epiphyseal growth plate osteoclasts preferentially phagocytose chondrocytes that are undergoing PCD. We injected biotin-labelled annexin-V (anx-V-biotin, an early marker of PCD) intravenously in young adult mice. After 30 min of labelling, long bones were recovered and the tissue distribution examined of anx-V-biotin-labelled cells in the growth plate using ABC-peroxidase histochemistry. Positive staining for anx-V-biotin was detected in hypertrophic chondrocytes still present in closed lacunae at some distance from the ossification front. At the ossification front, chondrocyte lacunae were opened and close contacts were seen between tartrate-resistant acid phosphatase-positive osteoclasts and hypertrophic cartilage cells. Osteoclasts were significantly more frequently in contact with anx-V-biotin-labelled chondrocytes than with unlabelled chondrocytes. Osteoclasts also contained labelled and unlabelled phagocytic fragments within their cytoplasm. We conclude that in the growth plate osteoclasts preferentially phagocytose hypertrophic chondrocytes that are dying, suggesting these dying cells may signal osteoclasts for their removal.