Combination therapy targeting integrins reduces glioblastoma tumor growth through antiangiogenic and direct antitumor activity and leads to activation of the pro-proliferative prolactin pathway

Abstract

Background: Tumors may develop resistance to specific angiogenic inhibitors via activation of alternative pathways. Therefore, multiple angiogenic pathways should be targeted to achieve significant angiogenic blockade. In this study we investigated the effects of a combined application of the angiogenic inhibitors endostatin and tumstatin in a model of human glioblastoma multiforme.

Results: Inhibitors released by stably transfected porcine aortic endothelial cells (PAE) showed anti-angiogenic activity in proliferation and wound-healing assays with endothelial cells (EC). Interestingly, combination of endostatin and tumstatin (ES + Tum) also reduced proliferation of glioma cells and additionally induced morphological changes and apoptosis in vitro. Microencapsulated PAE-cells producing these inhibitors were applied for local therapy in a subcutaneous glioblastoma model. When endostatin or tumstatin were applied separately, in vivo tumor growth was inhibited by 58% and 50%, respectively. Combined application of ES + Tum, in comparison, resulted in a significantly more pronounced inhibition of tumor growth (83%). cDNA microarrays of tumors treated with ES + Tum revealed an up-regulation of prolactin receptor (PRLR). ES + Tum-induced up-regulation of PRLR in glioma cells was also found in in vitro. Moreover, exogenous PRLR overexpression in vitro led to up-regulation of its ligand prolactin and increased proliferation suggesting a functional autocrine growth loop in these cells.

Conclusion: Our data indicate that integrin-targeting factors endostatin and tumstatin act additively by inhibiting glioblastoma growth via reduction of vessel density but also directly by affecting proliferation and viability of tumor cells. Treatment with the ES + Tum-combination activates the PRLR pro-proliferative pathway in glioblastoma. Future work will show whether the prolactin signaling pathway represents an additional target to improve therapeutic strategies in this entity.

Figure 1

Effects of conditioned medium from encapsulated PAE cells overexpressing ES and Tum. (A) Secretion of ES and Tum from stably transfected PAE cells was determined using Western blot analysis of culture supernatants after protein concentration with heparin sepharose for ES and Nickel Cam for Tum. (B) Transfected PAE cells were encapsulated in alginate/PLL as described. Phase contrast photomicrograph shows transfected cells in microbeads at a magnification of x10. (C) HUVECs cultivated in CM containing ES, Tum or ES + Tum showed reduced proliferation rates when compared with cells cultivated in CM from PAE WT cells. Bars represent mean values ± SE (n = 3); * on bars indicates significant differences vs. WT (p < 0.05); */** on connecting lines indicates significant differences between respective groups (p < 0.05 / p < 0.005). (D)Wound healing assay. CM containing ES, Tum or ES + Tum reduced migration of HDMECs. Wound closure data are normalized to results obtained with CM from PAE WT cells.

Figure 2

Conditioned medium containing ES + Tum reduced proliferation and induced apoptosis in G55 glioma cells in vitro. (A) G55 cells were cultivated in CM from PAE WT cells, and in CM containing ES or Tum or ES + Tum. Cell numbers were determined after a culture period of 24 and 48 hours. Bars represent the means ± SE (n = 3); * on bars indicates significant differences vs. WT; * on connecting lines indicates significant differences between respective groups (p < 0.05). (B) Phase contrast micrographs of G55 cells after a culture period of 24 hours in CM from PAE-WT cells or CM containing ES or Tum or ES + Tum. (C) Quantification of apoptotic G55 cells after cultivation in CM from PAE WT cells or CM containing ES or Tum or ES + Tum using flow cytometric analysis after FITC-conjugated annexin-V and PI staining.

Figure 3

Analysis of G55 tumors treated with ES, Tum and ES + Tum. Effect on tumor growth and tumor vessel density. (A) Subcutaneously grafted tumors were treated with ES, Tum and ES + Tum for 10 days, tumors were excised and tumor weights determined. Bars represent the mean values ± SE (n = 5); * on bars indicates significant differences vs. WT; * on connecting lines indicates significant differences between respective groups (p < 0.05). (B) Microvessel density was analysed by counting CD31-positive vessels in tumor tissue. Note that microvessel density is significantly reduced in ES- and ES + Tum-treated tumors, but not in Tum-treated tumors. Bars represent the means ± SE (n = 4-5); * significant differences vs. WT (p < 0.05). (C) Tumor tissue was analysed by H&E, TUNEL and immunohistochemical stainings for Ki67 and CD31. ES-, Tum- and ES + Tum-treated tumors showed large necrotic areas containing TUNEL-positive cells. Vital cells were restricted to thin cell layers around vessels (v) and at the outer margins of tumors. Control tumors (WT), in comparison, were highly proliferative and lacked large necrotic areas.

Figure 4

Expression analyses of prolactin receptor in tumor tissue. (A) Quantitative RT-PCR revealed a 2.5-fold upregulation of prolactin receptor mRNA expression in ES + Tum-treated tumors when compared to control tumors (B)Upper side: Immunostaining for prolactin receptor in control tumors (x10) and ES + Tum-treated tumors (left x10; right x40). Lower side: Immunostaining for prolactin receptor and cleaved cytokeratin (M30) in ES + Tum-treated tumors (x40).

Figure 5

Elevated levels of PRLR mRNA in glioma cells treated with ES + Tum. (A) Quantification of prolactin receptor mRNA expression revealed a 14-fold increase in G55 cells treated with CM containing ES + Tum when compared to G55 cells treated with CM from PA-WT cells. (B) Quantification of prolactin receptor mRNA expression in G55 cells treated with CM from PAE-WT, CM containing ES + Tum and cilengitide (CGT, 5 μg/ml) alone and in combinations. Bars represent the mean values ± SE from three independent experiments. (C) Immunofluorescence staining for PRLR in G55 cells after treatment with CM from PAE-WT and CM containing ES + Tum.

Figure 6

PRLR overexpression in glioma cells in vitro. (A) Verification of PRLR-overexpression in the glioma cell line G55 using Western blot analysis and RT-PCR (left). Quantification of PRLR and PRL mRNA expression in stably PRLR-transfected (G55 PRLR) and mock-transfected ((G55 3.1(−)) G55 cells (right). (B) Effect of PRLR overexpression on cell growth in the presence or absence of PRL and the Jak-2 inhibitor AG-490. Bars represent the mean values ± SE (n = 5-6); ** on connecting lines indicates significant differences between respective groups (p < 0.005). (C) PRLR- and mock- (3.1(−)) transfected G55 cells were treated with CM from PAE-WT or CM containing ES + Tum with and without the Jak-2 inhibitor AG-490. Bars represent the mean values ± SE (n = 5-6); * on connecting lines indicates significant differences between respective groups (p < 0. 05).