Osteopenia is a recognized complication of diabetes mellitus in humans and experimental animals. We recently found that tetracyclines prevent osteopenia in the streptozotocin-induced diabetic rat and that this effect was associated with a restoration of defective osteoblast morphology (Golub et al., 1990). The present study extends these initial ultrastructural observations by assessing osteoblast function in the untreated and tetracycline-treated diabetic rats. After a 3-week protocol, non-diabetic control and diabetic rats, including those orally administered a tetracycline, minocycline (MC), or a non-antimicrobial tetracycline analog (CMT), were perfusion-fixed with an aldehyde mixture; the humeri were dissected and processed for ultracytochemical localization of alkaline phosphatase (ALPase) and Ca-ATPase activities. Some rats from each experimental group received an intravenous injection of 3H-proline as a radioprecursor of procollagen, and the humeri were processed for light microscopic autoradiography. In addition, the osteoid volume in each experimental group was quantitatively examined by morphometric analysis of electron micrographs.

During the diabetic state, active cuboidal osteoblasts in the endosteum of control rats were replaced by flattened bone-lining cells that contained few cytoplasmic organelles for protein synthesis (Golgi-RER system), and active transport (mitochondria). Treating diabetic rats with MC, and even more so with CMT, appeared to “restore” osteoblast structure. During diabetes, bone-lining cells incorporated little 3H-proline or secreted little labeled protein and produced only a very thin osteoid layer. Tetracycline administration to the diabetics increased both the incorporation of 3H-proline by osteoblasts and their secretion of labeled protein toward the osteoid matrix, in a pattern similar to that seen in the non-diabetic controls. Intense alkaline phosphatase (ALPase) activity was cytochemically demonstrated along the plasma membranes of osteoblasts in the non-diabetic control rats, but was completely absent from the bone-lining cells in the diabetics. Similar to that described above, CMT therapy restored the ALPase activity in the diabetic osteoblasts and the effect of MC was less dramatic. The distribution and intensity of Ca-ATPase in the osteoblast-plasma membranes of the different groups of rats were similar to that of ALPase, except for the absence of detectable Ca-ATPase in the MC-treated diabetics. These results suggest that diabetes-induced osteopenia reflects, al least in part, impaired osteoblast structure and function and that tetracyclines, by a non-antimicrobial mechanism, may prevent this bone deficiency by normalizing these bone-lining cells.