Osteocytes are terminally differentiated long-lived cells and account for greater than 95% of the bone cell population. It has been established that osteocytes are connected through their highly developed dendritic network, which is necessary for the maintenance of optimal bone homeostasis. However, little is known on how osteocytes use the network to coordinate their cellular function and communication that requires energy and protein turnover. Here using super-resolution confocal imaging on both live and fixed osteocytes, we demonstrated conclusively that mitochondria are widely distributed and dynamically shared between osteocytes. Using confocal live cell imaging analysis we showed that inhibiting the contact between mitochondria and endoplasmic reticulum (ER) by the knockdown of MFN2 in osteocytes impedes the transfer of mitochondria suggesting the involvement of ER contact with mitochondria in the transfer process. Moreover, we showed that transport of mitochondria between osteocytes within the network enables rescue of osteocytes with dysfunction of mitochondria. Using the 3D tetraculture system with confocal imaging, we identify the transfer of mitochondria from healthy osteocytes enables recovery of mitochondria activities in osteocytes that devoid of mitochondrial DNA by ethidium bromide. The results indicated that when osteocytes are depleted of functional mitochondria, normal parental osteocytes can transfer mitochondria to these stressed osteocytes to provide them with energy. Collectively we show for the first time that the utilisation of mitochondrial transfer enables osteocytes to function with a network and coordinate their cellular activities in response to different energy demands.

Mechanical loading plays an essential role in both tendon development and degradation. However, the underlying mechanism of how tendons sense and response to mechanical loading remains largely unknown. SPARC, a multifunctional extracellular matrix glycoprotein, modulates cell extracellular matrix contact, cell-cell interaction, ECM deposition and cell migration. Adult mice with SPARC deficiency exhibited hypoplastic tendons in load-bearing zone. By investigating tendon maturation in different stages, we found that hypoplastic tendons developed at around postnatal 3 weeks when the mice became actively mobile. The in vitro experiments on primary tendon derived stem cells demonstrated that mechanical loading induced SPARC production and AKT/S6K signalling activation, which was disrupted by deleting SPARC causing reduced collagen type I production, suggesting that mechanical loading was harmful to tendon homeostasis without SPARC. In vivo treadmill training further confirmed that increased loading led to reduced Achilles tendon size and eventually caused tendon rupture in SPARC-/− mice, whereas no abnormality was seen in WT mice after training. We then investigate whether paralysing the hindlimb of SPARC-/− mice using BOTOX from postnatal 2 weeks to 5 weeks would delay the hypoplastic tendon development. Increased patellar tendon thickness was shown in SPARC-/− mice by reducing mechanical loading, whereas opposite effect was seen in WT mice. Finally, we identified a higher prevalence of a missense SNP in the SPARC gene in patients who suffered from a rotator cuff tear. In conclusion, SPARC is a mechano-sensor that regulates tendon development and homeostasis.