SPATIAL MPROBE: A MULTIMODAL IMAGING APPROACH FOR SPATIALLY RESOLVED GENE MECHANOREGULATION ANALYSIS DURING BONE FRACTURE HEALING

Local tissue mechanics play a crucial role in bone fracture healing. While the Mechanostat Theory by Frost describes an intuitive response for structural bone adaptation based on tissue strains, we asked how this regulation functions at the gene level. To test this, we have developed Spatial μProBe, a tool for multimodal correlation of in vivo micro-CT images with ex vivo histological sections from the same sample. Using micro-finite element analysis (micro-FE), mechanical strains can be computed in the local in vivo environment (LivE) and integrated with gene expression from spatial transcriptomics (ST) data, revealing changes in gene expression with local mechanical signals. Here, we analysed data from a 12-week-old female mouse that received a mid-diaphyseal femoral osteotomy, followed by individualised cyclic mechanical loading via an external fixator, from three weeks post-surgery onwards. Weekly in vivo micro-CT scans (10.5 μm, vivaCT 80) were performed until the mouse was euthanised (5 weeks post-surgery). Spatial μProBe facilitated co-registration of time-lapsed in vivo micro-CT images (10.5 μm) with a structurally intact paraffin section (FFPE, 5 μm thickness) processed with the 10x Genomics Visium ST protocol. Bone in the section was segmented manually and 2D–3D registered to the corresponding in vivo micro-CT scan. We used iStar to obtain super-resolution ST data at micro-CT resolution. Super-resolved ST data revealed the spatial expression for selected genes implicated in bone remodelling such as Col1a2 , encoding the pro-alpha2 chains of type I collagen, an abundant structural protein in bone and Ctsk , encoding cathepsin-K, a key lysosomal cysteine proteinase involved in bone resorption (Fig 1A). Micro-FE analysis was performed on binarised micro-CT scans (395 mgHA/cm 3 ), yielding the local mechanical signal as effective strain (μe) (Fig 1B). Col1a2 maintained high expression across all mechanical signals, with a sinusoidal-like shape, concordant with the potent anabolic stimulus induced by mechanical loading (Fig 1C). Conversely, Ctsk expression decreased with increasing mechanical signals and was strongly reduced in regions of high mechanical signals. This result agrees with the low likelihood of resorption events in areas of high strain and overall tighter regulation of bone resorption. Our results indicate that gene expression is controlled mechanically during fracture healing, suggesting that identifying functionally and therapeutically relevant strain targets may help improve healing outcomes. In conclusion, Spatial μProBe using super-resolution ST data can be used to investigate the complex mechanical regulation of gene expression, which can help improve our understanding of multiscale bone mechanobiology during fracture healing. For any figures or tables, please contact the authors directly.