Expression of Tax-interacting protein 1 (TIP-1) facilitates angiogenesis and tumor formation of human glioblastoma cells in nude mice

Abstract

Glioblastoma is the most common and fatal type of primary brain tumors featured with hyperplastic blood vessels. Here, we performed meta-analyses of published data and established a correlation between high TIP-1 expression levels and the poor prognosis of glioblastoma patients. Next, we explored the biological relevance of TIP-1 expression in the pathogenesis of glioblastoma. By using orthotopic and heterotopic mouse models of human glioblastomas, this study has characterized TIP-1 as one contributing factor to the tumor-driven angiogenesis. In vitro and in vivo functional assays, along with biochemical analyses with microarrays and antibody arrays, have demonstrated that TIP-1 utilizes multiple pathways including modulating fibronectin gene expression and uPA protein secretion, to establish or maintain a pro-angiogenic microenvironment within human glioblastoma. In conclusion, this work supports one hypothesis that TIP-1 represents a novel prognostic biomarker and a therapeutic target of human glioblastoma.

Fig. 1

Elevated TIP-1 expression levels correlate with the poor prognosis of human glioblastomas. (A) Relative TIP-1 expression levels in human malignant gliomas of various stages upon WHO grading standard. The data distribution was presented as box plot. Note the correlation of the TIP-1 expression levels with the pathological grades of tumors. (B) Kaplan-Meier survival curve to show survival probability of glioblastoma (WHO grade IV) patients after diagnosis is stratified by the TIP-1 expression levels (log-rank analysis, p=0.0274).

Fig. 2

TIP-1 knockdown within human glioblastoma cell lines delays tumor formation and impairs angiogenesis in nude mice. D54 or U87 cells were genetically modified with recombinant lentiviruses expressing luciferase and TIP-1-targeted or a control shRNA before the intracranial inoculation. TIP-1 expression in the transfected cells was verified by western blot analyses with a specific antibody (inserts in B & C). Tumor formation and growth were monitored with bioluminescence imaging (BLI) upon luciferase expression. (A) Representative BLI data collected at day 21 post D54 cell inoculation, quantitative measurements of tumor growth were shown upon BLI signal intensity. (B) D54 and (C) The graphs show the time course of tumor formation and growth of D54 (B) and U87 (C) cells in nude mice (n = 5 for each group). (D) Growth curves of D54 tumors in a subcutaneously implanted xenograft model (n=5 for each group). A dashed line indicates tumor size as of 50 mm³. (E) Tumor-associated blood vessels were visualized with vWF staning (brown) and quantified upon images. Tumors were resected at day 12 post inoculation in a subcutaneously implanted xenograft model. Hematoxylin (blue) was used for counterstaining. Quantification was based upon more than 12 independent microscopic fields for each group. Shown are mean +/− SD. * p<0.05 compared to the shRNA control by the Student’s t-test.

Fig. 3

TIP-1 knockdown impairs production of angiogenic factors within D54 cells. (A) Matrigel-plug assays. A half million D54 cells with or without TIP-1 knockdown were mixed with matrigel for subcutaneous implantation in nude mice. FITC-dextran was injected into the tail vein on day 8 after the matrigel plug implantation. Tumors were resected for examination with a confocal fluorescence microscope. Shown are representative images and quantitative measurement (ImageJ) of the FITC signals from at least 12 microscopic fields. * p<0.05 compared to the shRNA control. (B) Angiogenic activity of pre-conditioned medium from monolayer culture of D54 cells was analyzed with a Boyden chamber-based HUVEC migration assay. Shown are the means ± SDs of three experiments in which HUVECs migrated through the porous chambers. (C) HUVEC tubule formation assay using pre-conditioned medium from monolayer culture of D54 cells. Shown are the quantitative measurements of capillary connections with more than 3 branches. Bars represent the means ± SDs of three independent experiments. * p<0.05 compared to the shRNA control by the Student’s t-test.

Fig. 4

TIP-1 knockdown suppresses secretion of angiogenic factors from glioblastoma cells. (A) Angiogenic factors in the D54 pre-conditioned medium were detected with a Human Angiogenesis Antibody Array kit. Shown are representative images of three independent experiments and quantitative measurements of the mean signal intensity of each spot with ImageJ. Identity of each arrayed antibodies is indicated in the legend at right. * p<0.05. (B) uPA in the pre-conditioned medium was measured with an ELISA kit, the optical density was converted to uPA concentration based on the titration results with a standard control. Shown are the means ± SDs from three experiments in D54 (left) and U87 (right) cells with or without TIP-1 knockdown. * p<0.05 compared to the controls by the Student’s t-test.

Fig. 5

TIP-1 knockdown reduces fibronectin expression in glioblastoma cells. (A) Heatmap showing the expression levels of some angiogenesis-related genes within D54 cells with or without TIP-1 knockdown. TIP-1 was included to show the efficiency of TIP-1 knockdown. (B) Quantitative RT-PCR analyses of FN1, SPP1 and IGFBP3 gene expression within D54 cells. Shown are fold changes at mRNA level of each gene, compared to that in the control cells (set to a value of 1). Means ± SDs from three indepdendent experiments. * p<0.05 compared to the shRNA control by the Student’s t-test. (C) Western blot analyses and semi-quantification of FN1, IGFBP3 and SPP1 protein levels in human glioblastoma cell lines D54 and U87. TIP-1 and actin were blotted as controls. Intensity of the protein bands was analyzed with ImageJ and normalized to the loading control (actin). The expression level of each protein was normalized upon those in the control cells (set to a value of 1).

Fig. 6

FN1 is involved in the TIP-1-regulated angiogenesis in human glioblastomas. (A) D54 cells with or without TIP-1 knockdown were mixed with 25 μg/ml of recombinant FN1 protein or PBS control for the matrigel-plug assays. Tumor blood vessels within the matrigel-plugs were detected and quantified at day 8 after plug implantation as described in Materials and Methods. p<0.05, n=5. The Student’s t-test. (B) FN1 expression within D54 cells was down-regulated with two validated siRNAs against human FN1 transcripts. The efficiency of siRNA was detected with western blot analyses with specific antibodies (lane 1, control siRNA; lane 2, FN1 siRNA #1; and lane 3, FN1 siRNA #2) as shown in the insert. The transfected cells were used for the matrigel-plug assays 72 hours post-transfection. Shown are representative images and quantitative measurements of fluorescence intensity (ImageJ). * p<0.05 compared to controls by the Student’s t-test.

Fig. 7

B>. Schematic illustration of TIP-1’s roles in the formation and growth of glioblastoma. Solid lines indicate documented pathways, while dashed lines show putative pathways to be explored.