In Vivo Imaging Model of Multiple Myeloma and Its Cellular Interaction with the Bone Marrow Microenvironment

BACKGROUND: Imaging animal models that offer serial measurement of systemic tumor progression, such as the GFP+ or bioluminescence MM model, have been limited to low resolution, gross measurements of tumor progression that are insufficient to detect individual cells, and their interaction with their microenvironment. Therefore, the need exists for development of sensitive, high resolution three-dimensional imaging methods that identify the dynamic changes that occur during tumor initiation and progression. We here show the use of in vivo fluorescence confocal microscopy to follow MM tumor initiation and progression at the cellular level using stably GFP-transfected MM1S cells in a xenograft model of MM.

METHODS: 5 × 10 6 MM1S-GFP-Luc cells were injected into the tail veins of non-irradiated SCID/Beige male mice. MM cell growth in the marrow of the calvarial bone was analyzed using in vivo flow cytometry and fluorescence confocal microscopy, as previously described (Sipkins et al 2005). High-resolution images with unprecedented cellular detail were obtained through the intact mouse skull at depths of up to 250μm. To visualize the bone marrow vasculature the mice were injected with a blood pool marker (Angiosense 680 or 750) immediately before imaging, and to delineate the surface of calvarial bone, a fluorescent hydroxyapatite tag (Osteosense) was used. The validity of the imaging data was established by sacrificing select mice, and analyzing the previously imaged tissues by standard histologic and immunohistologic techniques. After MM tumors became established in the fourth week following injection, 1 mg/kg Bortezomib was administered twice weekly to a subset of the mice, these were imaged following treatment along with controls that were not treated. For all mice imaged, the number and areas of the skull where GFP+ MM cells were found were recorded. Confirmation of homing and tumor progression was also performed using CD138+ selected primary tumor cells.

RESULTS: Using this model, we were able to detect and monitor individual GFP+MM cells within the bone marrow microenvironment. We demonstrate that MM.1 S cells and primary CD138+ cells exit the systemic circulation within one hour of injection, followed by specific rolling and adhesion to the vasculature of the bone marrow microenvironment. Within 4 days post after injection, the MM cells were fully engrafted along the bone marrow sinusoids, which were surrounded by bisphosphonate-rich bone structures including ostoeoblasts. Within the second week, loose clusters of a few cells began to form around the blood vessels. Growth and expansion appeared to be closely associated with the vasculature. Tumor growth dramatically increased in the third week following cell injection when areas of the parasagittal regions became completely involved with MM cells. In contrast, standard bioluminescence imaging performed concurrently detected tumor initiation only at 4 weeks post-injection, indicating that confocal microscopy is a much more sensitive technique in detecting early tumor proliferation. Imaging of bortezomib-treated mice demonstrated that tumor size and density was reduced in the skull, but even more dramatically the number of sites containing GFP+MM was greatly reduced.

CONCLUSIONS: Our imaging model differs from other models due to its unprecedented resolution. Therefore it is particularly useful for following small numbers of tumor cells either early in disease progression or after therapeutic treatment. This model offers a more sensitive spatial and temporal live imaging of MM cells in the BM microenvironment and can be used to explore the dynamic interaction of MM with different structures and environments of the BM. We anticipate that this model will allow for a better understanding of the biologic effects of therapeutic agents on the growth of MM cells within the bone marrow niches.