Poster R14 - Characterizing the limits of 3-photon microscopy as a tool for deep imaging of whole, cortical bone tissue

Introduction: Osteocytes have emerged as the primary mechanosensors in bone with a rising need to study their mechanobiology within the context of their in vivo microenvironment. These cells form dense networks interconnected with slender canaliculi. However, many optical techniques struggle to penetrate the dense and highly scattering mineralization within bone. Two-photon (2P) microscopy can successfully observe osteocyte molecular signaling in vivo in the third metatarsal (MT3) of mice but remains limited by achievable imaging depth. Three-photon (3P) microscopy is a novel imaging platform that offers increased depth penetration over 2P fluorescence imaging. In addition, third harmonic generation (THG) reveals material interfaces without the need for exogenous labels, allowing for the visualization of the interface between the bone and lacuno-canalicular network (LCN). Indeed, 3P has been used to observe in vivo osteocytes and canaliculi within mouse calvaria, penetrating to depths that reach the marrow where bone lining cells were visible. Nonetheless, further characterization is needed to establish 3P imaging in long bones, which have thicker cortices and are of greater biomechanical interest. In this study, we hypothesized that 3P microscopy has better imaging depth and quality compared to 2P in mouse long bones ex vivo.

Materials and

Methods: All procedures were approved by the Cornell University IACUC. Bilateral third metatarsals (MT3s), tibiae, femora, and humeri of 16-20 week old mice expressing osteocyte-targeted GCaMP6f were dissected, fixed in zinc formalin, and soaked overnight in 1mM CaCl2 to maximize GCaMP6f fluorescence. Images were collected using a Monaco laser (Coherent) with a 25x objective (Olympus) at 920nm for 2P excitation and 1320nm for 3P excitation. Z-stacks at 512x512 pixels were taken at the diaphysis while power was exponentially adjusted through depth. Image collection continued through bone depth at 1μm steps and averaged at 5 frames until signal was no longer visually observed. Stacks were cropped in FIJI to a 200μm wide rectangle centered on the bone to avoid variations in depth due to the curved periosteal surface, and slices beyond the endocortical surface in the MT3 were excluded. Automated 3D segmentation was used to create ROIs around cells/lacunae, and intensities were normalized to the background. ROIs with surface areas smaller than 250 pixels and greater than 2,250 pixels were empirically determined to not accurately capture cells and excluded. Signal-to-noise ratio (SNR) was calculated as the intensity divided by the standard deviation of the background, while signal-to-power ratio (SPR) was calculated as the natural log of the signal intensity divided by the power. Unpaired t-tests, kolmogorov-smirnov tests, and linear regressions were calculated in R.

Results, Conclusions, and Discussions: There was no significant difference in the average maximum depth at which fluorescent signal was collected by 2P or 3P, and only in the femur were fewer cell ROIs captured by 3P (Fig A-B). Unexpectedly, THG did not reach depths equivalent to fluorescent signal but collected similar lacunae ROI quantities across all bones (Fig C-D). The distribution of ROIs through the depth in 2P and 3P was different in the tibia, femur, and humerus, with 2P showing a moderate to high right skew and 3P shifted deeper (Fig E-H). SNR and SPR showed stable or increasing trends over depth in 3P while 2P generally decreased (Fig I-L). Interestingly, 3P exceeded 2P performance below ~60μm in the femur and humerus. THG signal was as bright or brighter than 3P fluorescent signal with stable SPR in all bones except the humerus (Fig M-P).
Here we report the establishment of a new multiphoton imaging approach, three-photon microscopy, for studying embedded osteocytes in mouse long bones. While similar quantities of data are collected by 2P and 3P fluorescence, differences in data quality become apparent as cortical depth increases with 3P performing similarly or greater than 2P at distinctly capturing objects of interest. THG brightly captures lacunae through depth and adds additional valuable data on the osteocyte microenvironment. These data suggest 3P may be used to advance in vivo observation of osteocytes.

Acknowledgements (Optional):