Supplemental Figure Legends

Figure 1. Analysis of integrin protein expression on a2 integrin shRNA-transduced PCa cells. A-G) Flow cytometry. Control shRNA-transduced and a2 integrin shRNA-transduced LNCaPcol-luc cells were evaluated for the surface expression of the indicated integrin subunits by flow cytometry as described in the Material and Methods section. H) Quantitative PCR for b1 integrin transcripts. Total RNA was isolated from LNCaPcol-luc control shRNA-transduced and a2 integrin shRNA-transduced cells and evaluated for changes in b1 integrin mRNA transcripts by quantitative PCR as described in the Material and Methods section. Shown is the mean ± the standard deviation of duplicate experiments.

Figure 2. Blocking a2 integrin reduces PCa establishment within the bone. Intratibial injection. Control shRNA-transduced or a2 integrin shRNA-transduced LNCaPcol-luc cells (5x105 cells/50 ml) were directly injected into the right proximal tibia of anesthetized male SCID mice, 12 mice/group. Tumors were allowed to grow for 9 weeks. Tumor burden was measured using bioluminescent imaging following the injection of luciferin. Shown are the corresponding bioluminescent images, x-rays, and histological sections of the mouse within each group that was closest to the median of the bioluminescent data shown in Figure 3A. Arrows indicate osteolytic activity within injected tibiae. Note the replacement of trabeculae (arrowhead) with tumor (T) within the tibia injected with control shRNA-transduced compared to a2 integrin shRNA-transduced cells. Also reported is the bone mineral density (BMD) of each tibia as measured by dual-energy X-ray absorptiometry (DEXA) using a pDEXA Sabre scanner (Orthometrix, Inc., White Plains, NY).

Figure 3. a2 integrin knock-down does not affect attachment dependent growth in vitro.

Triplicate wells on a 96-well tissue culture plate were coated with 1 mg/cm2 human type I

collagen, fibronectin, laminin, or vitronectin for 2 hours at 37°C. Control shRNA-transduced and a2 integrin shRNA-transduced LNCaPcol-luc cells were plated to each of four, coated plates in triplicate (1.5x103 cells/0.2ml/well). The total number of viable cells was determined every 24 hours for four days with the addition of 1mM MTT. The data is presented as the mean ± standard deviation of triplicate wells from a representative experiment.

Figure 4. Analysis of PCa experimental metastases formed following intracardiac injection. Control shRNA-transduced or a2 integrin shRNA-transduced LNCaPcol-luc cells were injected into the left cardiac ventricle (2x105 cells/0.1 ml), 11 mice per group. Tumors were allowed to grow for 9 weeks and tumor burden was measured using bioluminescent imaging following the injection of luciferin. Shown are representative bioluminescent images and corresponding histological sections of the hepatic metastases produced by control- or a2 shRNA-tranduced cells.

Figure 5. Integrin a2β1 expression correlates with skeletal metastasis in PCa patients. Tissue microarrays constructed from samples of non-neoplastic prostate tissue, BPH, primary PCa lesions, and soft tissue and skeletal metastases were stained for α2 integrin expression using routine immunohistochemistry as outlined in the Methods section. Shown are representative images of each tissue type. Brown color indicates a2 staining.

2