Fig. 3

Zoledronic acid (ZA), but not Denosumab, reduces metastatic load and increases bone formation in humanized tissue-engineered bone constructs (hTEBCs). a Ex vivo bioluminescence imaging (BLI) revealed a significant reduction of the metastatic load in hTEBCs by ZA compared with Denosumab and PBS. The Tukey method was used to determine outliers (shown as individual data points). One data point for Denosumab and two for ZA were excluded for calculation of P-values using a UNIANOVA model (**P ≤ 0.01, ***P ≤ 0.001; n = 11 for Denosumab; n = 8 for ZA; n = 10 for PBS). BLI images of included hTEBCs are shown to the right of the graph. b Fold change of the bone volume (BV) at the hTEBC between week 8 (before cancer cell injection) and 17. hTEBCs in mice treated with Denosumab or PBS maintained a constant BV after cancer cell injection. Only animals treated with ZA showed an increase in BV after PC3-Luc injection of ~2.5-fold, however without statistical significance. c The histology (H&E) overview section (left panel) confirmed a PC3-Luc metastatic lesion within the hTEBC (black square). In the zoom in (right panel) the bone marrow compartment (black arrowheads) and mineralized tissue (asterisks) resembling a human bone niche with trabecular and cortical bone are apparent. d Immunohistochemistry (IHC) staining for human LaminAC confirmed the human origin of metastases in hTEBCs (left panel). IHC staining for CD44 confirmed that the lesion originated from prostate cancer cells (right panel). e Tartrate-resistant acid phosphatase (TRAP) staining of the hTEBC revealed an invasive line of TRAP-positive cells of the osteolytic PC3-Luc lesion in the PBS and the Denosumab group (TRAP1 respectively, white arrow heads). In addition, physiological bone turnover was observed at the interface of the human mineralized tissue within the hTEBC for all three experimental groups (TRAP2 for PBS and Denosumab group, white arrow heads). f Quantification of the TRAP-positive tissue area for each treatment group showed a significant reduction of osteoclast activity after treatment with ZA, but not for the Denosumab and PBS group. P-values were calculated using a UNIANOVA model (*P ≤ 0.05; n = 15 for Denosumab; n = 16 for ZA; n = 13 for PBS)