3D visualization and characterization of the early bone metastatic niche: from cancer cell homing to small osteolytic lesions

Breast cancer is one of the main cancer-associated causes of death in women. While a primary tumor can often be treated well, metastases have a very poor prognosis and the spread to bones is among the most frequent forms of metastasis with very poor outcome. While the cellular involvement of late-stage metastasis in the bone has been studied frequently, early metastatic lesions and associated structural and biophysical parameters, as well as the dynamic bone (re)modeling behavior are not well understood. We hypothesize, that such structural and biophysical changes in the bone microenvironment, influence the establishment and progression of breast cancer induced osteolytic lesions. Hence, we aim to develop an experimental mouse model of breast cancer bone metastasis to detect and characterize early osteolytic lesions in 3D. In order to study the dynamic bone (re)modeling behavior in a pathological model, a physiological baseline has to be established. In vivo microCT-based time-lapse morphometry is a powerful tool to study temporal and spatial changes in bone (re)modeling. Here we present an advancement of the method to detect and quantify site-specific differences of bone (re)modeling in 12-week-old female BALB/c nude mice. We establish new bone surface interface parameters and show how they are affected by bone curvature. Significant differences in bone (re)modeling baseline parameters between metaphysis and epiphysis, as well as distal femur and proximal tibia, for both cortical and trabecular bone, are described, with important implications for disease models. This baseline of physiological bone (re)modeling using our advanced microCT-based time-lapse morphometry method is then used to study changes in the bone dynamics caused by breast cancer cell bone metastasis. For this we inject mice with breast cancer cells and monitor the bone (re)modeling to detect pathological changes. We show that tumor-injected animals without osteolytic lesions have significantly higher parameters for newly mineralized bone in the trabecular region, compared to healthy control mice and with similar trends for cortical bone. This indicates an influence of cancer cells on the bone (re)modeling even in the absence of detectable lesions and a possible establishment of a pre- or antimetastatic niche. In order to study early osteolytic lesions caused by breast cancer cells in the bone, we develop an eroded bone patch analysis tool. This new mathematical tool allows us to detect and quantify cortical osteolytic lesions already two weeks after cancer cell injection, clearly distinguishing the pathological and physiological eroded bone patches. In addition, we visually identify lesions in the primary spongiosa of trabecular bone, sitting directly under the mineralized growth plate, already two weeks after cancer cell injection. MicroCT-based time-lapse morphometry allows us to describe for the first time three different types of early osteolytic lesions in the bone: 1) cortical lesions initiating at the periosteum, 2) cortical lesions at the endocortial site with additional trabecular erosion and 3) trabecular lesions in the primary spongiosa at the growth plate. We then use our in vivo results to study the homing of cancer cells in the bone using light-sheet fluorescence microscopy and confocal laser scanning microscopy, as well as the tissue changes caused by early osteolytic lesions with the help of backscattered electron microscopy and advanced confocal laser scanning microscopy. We study the size and location of cancer cells in 3D (intact) bones after optical clearing with the help of light-sheet fluorescence microscopy, providing 3D quantification of the homing of cancer cells in the bone marrow and bone surrounding tissue. Within the bone marrow, cancer cells home as small cell clusters close to the endocortical bone, but with no apparent preference for different bone compartments. Further analysis of cancer cell clusters in the marrow revealed that a significant fraction is not proliferating. Additionally, cancer cell clusters have a strong tendency to home in fibronectin-rich areas, providing new implications for the structural features of the cancer cell niche. We last perform a multiscale analysis of the early metastatic lesions with various imaging techniques and are able to show the changes in the mineralized tissue, as well as the organic collagen matrix. To sum up, we use microCT-based time-lapse morphometry to study the dynamics and onset of bone metastasis, including a baseline to differentiate from physiological bone (re)modeling. We quantify changes in pathological bone (re)modeling in the absence of detectable osteolytic lesions. Further, we introduce a new tool to detect and quantify early osteolytic lesions in cortical bone. In addition, we visually detect trabecular lesions and are able to classify three different types of lesions in cortical and trabecular bone. Using advanced ex vivo multimodal tissue analyses, we describe the homing of cancer cells to the bone marrow in 3D and characterize the bone microenvironment in early osteolytic lesion. Our work gives important 3D information and new perspectives on various states of cancer research including the debate on pre- or antimetastatic niches, homing and the onset of metastasis in the long bones