Comparison of ex vivo bioluminescence imaging, Alu-qPCR and histology for the quantification of spontaneous lung and bone metastases in subcutaneous xenograft mouse models.

Bioluminescence imaging (BLI) is a non-invasive state-of-the-art-method for longitudinal tracking of tumor cells in mice. The technique is commonly used to determine bone metastatic burden in vivo and also suitable ex vivo to detect even smallest bone micro-metastases in spontaneous metastasis xenograft models. However, it is unclear to which extent ex vivo BLI correlates with alternative methods for metastasis quantification. Here, we compared ex vivo BLI, human DNA-based Alu-qPCR, and histology for the quantification of bone vs. lung metastases, which are amongst the most common sites of metastasis in prostate cancer (PCa) patients and spontaneous PCa xenograft models. Data from 93 immunodeficient mice were considered, each of which were subcutaneously injected with luciferase/RGB-labeled human PCa PC-3 cells. The primary tumors were resected at ~ 0.75 cm³ and mice were sacrificed ~ 3 weeks after surgery and immediately examined by ex vivo BLI. Afterwards, the right lungs and hind limbs with the higher BLI signal (BLIHi bone) were processed for histology, whereas the left lung lobes and hind limbs with the lower BLI signal (BLILo bone) were prepared for Alu-qPCR. Our data demonstrate remarkable differences in the correlation coefficients of the different methods for lung metastasis detection (r ~ 0.8) vs. bone metastasis detection (r ~ 0.4). However, the BLI values of the BLIHi and BLILo bones correlated very strongly (r ~ 0.9), indicating that the method per se was reliable under identical limitations; the overall level of metastasis to contralateral bones was astonishingly similar. Instead, the level of lung metastasis only weakly to moderately correlated with the level of bone metastasis formation. Summarized, we observed a considerable discrepancy between ex vivo BLI and histology/Alu-qPCR in the quantification of bone metastases, which was not observed in the case of lung metastases. Future studies using ex vivo BLI for bone metastasis quantification should combine multiple methods to accurately determine metastatic load in bone samples.

Spontaneous metastasis xenograft models link CD44 isoform 4 to angiogenesis, hypoxia, EMT and mitochondria-related pathways in colorectal cancer.

Hematogenous metastasis limits the survival of colorectal cancer (CRC) patients. Here, we illuminated the roles of CD44 isoforms in this process. Isoforms 3 and 4 were predominantly expressed in CRC patients. CD44 isoform 4 indicated poor outcome and correlated with epithelial-mesenchymal transition (EMT) and decreased oxidative phosphorylation (OxPhos) in patients; opposite associations were found for isoform 3. Pan-CD44 knockdown (kd) independently impaired primary tumor formation and abrogated distant metastasis in CRC xenografts. The xenograft tumors mainly expressed the clinically relevant CD44 isoforms 3 and 4. Both isoforms were enhanced in the paranecrotic, hypoxic tumor regions but were generally absent in lung metastases. Upon CD44 kd, tumor angiogenesis was increased in the paranecrotic areas, accompanied by reduced hypoxia-inducible factor-1α and CEACAM5 but increased E-cadherin expression. Mitochondrial genes and proteins were induced upon pan-CD44 kd, as were OxPhos genes. Hypoxia increased VEGF release from tumor spheres, particularly upon CD44 kd. Genes affected upon CD44 kd in xenografts specifically overlapped concordantly with genes correlating with CD44 isoform 4 (but not isoform 3) in patients, validating the clinical relevance of the used model and highlighting the metastasis-promoting role of CD44 isoform 4.