Collagen mineralization is required for bone and dentin formation, but the molecular mechanism of that process is not completely understood. Using an in vitro model based on mouse dental tissues, we show that glycosaminoglycans (GAGs) promote collagen mineralization in mineralized tissues, rather than inhibiting mineralization of the ligament. Moreover, the mineralization promoting effect of GAGs is larger than for phosphoproteins, the most frequently studied controlling factor in crystal nucleation and growth. Our model enabled the study of molecular factors on collagen mineralization independently, without changing the expression pattern of other factors, and under conditions as close to native tissue as possible. Evidence that matrix GAGs promote collagen biomineralization has important implications for understanding control of mineralization in connective tissue more generally.
Mammalian teeth are attached to the jawbone through an exquisitely controlled mineralization process: unmineralized collagen fibers of the periodontal ligament anchor directly into the outer layer of adjoining mineralized tissues (cementum and bone). The sharp interface between mineralized and nonmineralized collagenous tissues makes this an excellent model to study the mechanisms by which extracellular matrix macromolecules control collagen mineralization. While acidic phosphoproteins, localized in the mineralized tissues, play key roles in control of mineralization, the role of glycosaminoglycans (GAGs) is less clear. As several proteoglycans are found only in the periodontal ligament, it has been hypothesized that these inhibit mineralization of collagen in this tissue. Here we used an in vitro model based on remineralization of mouse dental tissues to determine the role of matrix GAGs in control of mineralization. GAGs were selectively removed from demineralized mouse periodontal sections via enzymatic digestion. Proteomic analysis confirmed that enzymatic GAG removal does not significantly alter protein content. Analysis of remineralized tissue sections by transmission electron microscopy (TEM) shows that GAG removal reduced the rate of remineralization in mineralized tissues compared to the untreated control, while the ligament remained unmineralized. Protein removal with trypsin also reduced the rate of mineralization, but to a lesser extent than GAG removal, despite a much larger effect on protein content. These results indicate that GAGs promote mineralization in mineralized dental tissues rather than inhibiting mineral formation in the ligament, which may have broader implications for understanding control of collagen mineralization in connective tissues.