Osteopontin gene expression determines spontaneous metastatic performance of orthotopic human breast cancer xenografts

Abstract

A major problem in the therapeutic management of cancer is the growth of metastases in distant organs, but the genes orchestrating the process need to be identified for the rational design of new treatment. Here, we provide decisive experimental evidence demonstrating the causal involvement of a specific gene, osteopontin (OPN), in the pathogenesis of metastasis by human breast cancer cells and implicating some of its probable partners. Stable long-term depletion, or up-regulation, of OPN gene expression in a matched, isogenic pair of human breast cancer cell lines of differing metastatic proficiency reproducibly changed their ability to colonize distant organs. OPN down-regulation was achieved by transduction of the metastatic line with a DNA construct encoding a small hairpin RNA in a vector labeled with red fluorescent protein and resulted in a marked reduction of metastatic load (P < 0.01). Up-regulation of OPN in the negligibly metastatic line, with a green fluorescent protein-marked retroviral vector containing OPN cDNA driven by a strong promoter, resulted in heavy colonization of the lungs and lymph nodes (P < 0.005). The reciprocal changes in behavior of these matched cell lines cross-corroborate each other. Concomitant changes were seen in the expression of other metastasis-related genes in both modulated lines. The data indicate that therapeutic targeting of tumor OPN molecules could reset metastatically relevant gene networks, resulting in clinical benefit.

Figure 1

a: Human OPN transcript levels in control and test cell lines measured by qPCR. b: Human OPN secretion levels into medium by control and test cell lines measured by ELISA. c: Morphology of unmodified 4M3 cells. d: Morphology of transduced 4OPsi cells. Note increased tendency to spindle-shape. e: Transduced 4OPVO cells after flow sorting. Note all cells fluoresce, but intensity varies. f: Transduced 4OPsi cells after flow sorting. Note spindle shape and variation in fluorescence intensity although all cells are labeled. g: Transduced 2C5GFP (control) cells after G-418 selection. Note uniform intensity of fluorescence. h: Transduced 2OPUR cells after sorting. Fluorescence is uniform, and note that there are no obvious morphological differences from control cells.

Figure 2

a: Growth curves of control and genetically modified cells in vitro. b: Attachment of control and modified cells to fibronectin- and albumin-coated plastic. c: Motility of control and modified cell lines tested by scratch-wound healing assay. d: Chemotaxis of cell lines toward fibronectin and bovine serum albumin. e: Invasion of control and test cell lines through Matrigel matrices.

Figure 3

a: Growth rates of tumors generated by intramammary inoculation of control and modified cell lines. b: Human OPN levels in tumors made by control and test cell lines (ELISA). c: Human OPN levels in blood of corresponding tumor-bearers measured by ELISA. d: Human uPA levels in tumors generated by control and modified cell lines (ELISA). e: Human MMP-2 levels in blood of animals bearing tumors generated by control and test cell lines (ELISA).

Figure 4

Necropsy results. a and d: Survey views of whole animal (a) and lungs (d) of 4OPVO animal show red fluorescent primary tumor and many metastases. Arrows in a show tumor deposits in lymph nodes. b and e: Survey views of animal inoculated with 4OPsi cell line show reduction in metastases in lymph nodes (b) and lungs (e). White arrow in b shows weakly fluorescing normal gall bladder. Primary tumor in mammary gland fluoresces red. c: Survey view of animal inoculated with 2OPUR cells shows GFP-labeled metastases in abdominal, thoracic, axillary, and cervical lymph nodes (red arrows). P, primary tumor. f: Increased metastases in lungs of 2OPUR animals compared with controls (see j). g: Closeup of f to show metastases and many intervening scattered cells in the lungs (×50). h: Lymph node from 2OPUR animal shows complete colonization of node and obstruction of adjacent lymph vessels (arrows) with tumor cells (×40). i: Survey view of the body of a control 2C5GFP animal shows an occasional positive lymph node and large green fluorescent primary tumor in mammary gland. j: Lungs of control 2C5GFP animal with occasional small metastases in lungs. Only deposits above 1 mm (see scale included) are metastases. k: Lungs of control 4M3-inoculated animal show the vigorous metastatic capability of the parental cell line.

Figure 5

a: RT-PCR products using murine-specific (m) and human-specific (h) OPN primers demonstrate the absence of mouse OPN in normal murine virgin mammary glands. NM2C5 tumor was added as positive control. (NTC, no template negative control.) Mouse glyceraldehyde-3-phosphate dehydrogenase amplicons were used as loading control. b: Human and murine OPN in xenogeneic tumors analyzed by species-specific qPCR. The tumor cells induced the non-neoplastic gland to secrete murine OPN, but the majority of the secreted protein is human. c: ELISA analysis showing that genetically increased OPN secretion is accompanied by decreased MMP-8 secretion.