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

Human glioblastoma multiforme (GBM) is the most common and malignant type of brain tumors. Current treatment options such as surgical intervention, radiation therapy or cytotoxic chemotherapy do not significantly improve the median survival beyond approximately 12 to 18 months for patients with GBM [1, 2]. Therefore, the identification and the development of novel and more efficient therapeutic approaches remain a crucial task for this disease.

Since GBM is characterized by particularly high levels of neovascularization, a therapeutic strategy based on angiogenic blockade appears to be promising. Actually, a number of strategies targeting new blood vessel formation have shown some success in preclinical models of GBM [3, 4] and several clinical trials with anti-angiogenic agents are ongoing [5]. An important feature of angiogenesis is the interaction of endothelial cells (EC) with surrounding extracellular matrix (ECM). Integrin binding mediates cell adhesion of ECs to surrounding ECM and regulates their survival, growth and mobility [6]. Integrins and αVβ5 are predominantly expressed in proangiogenic ECs [7, 8] and especially integrin αVβ3 has been found to be upregulated in ECs of GBM tumors [9, 10]. Cilengitide, a cyclic pentapeptide mimicking the Arg-Gly-Asp (RGD) binding site of integrin ligands, was identified as a potent and selective integrin antagonist that interfered with binding of ECM components to αVβ3 and αVβ5 integrins [11]. In preclinical models cilengitide had synergistic therapeutic effects with radioimmunotherapy in breast cancer and orthotopic brain tumor models [12, 13]. However, expression of αVβ3 and αVβ5 integrins is not restricted to activated ECs. Both integrins are also in brain tumor cells [14‐16]. In fact, we have recently shown that cilengitide inhibits integrin-dependent signaling and induces apoptosis not only in endothelial but also in glioma cells thereby explaining the profound activity of integrin inhibitors in this disease [17]. These data suggest that anti-angiogenic molecules directed towards integrins may have a multi-targeting effect on both endothelial and glioma cells.

An additional aspect to be considered for the design of novel therapeutic strategies against GBM is the ability of these tumors to escape anti-angiogenic monotherapy [18, 19]. Therefore, it might be necessary to target multiple pro-angiogenic pathways in order to achieve significant anti-tumorigenic effects.

Here, we studied two angiogenic inhibitors targeting different angiogenic pathways, endostatin (ES) and tumstatin (Tum), and evaluated the anti-tumorigenic activity of the individual factors and a combination of both factors in an in vivo model of GBM. ES has been reported to interfere with integrin α5β1 and VEGFR-2 in ECs, while Tum binds αvβ3 and αVβ5 integrins and induces apoptosis in ECs [20, 21]. In addition, microarray analysis of tumor tissue was performed to identify activation of alternative pro-tumorigenic signalling pathways in tumor cells.

Angiogenesis plays a central role in tumor growth and metastasis. Since GBM tumors are highly vascularised, therapeutic strategies based on angiogenic blockade are particularly attractive for this entity. However, it has been observed that initial responses to anti-angiogenic therapy are frequently followed by tumor progression resulting in only limited survival advantage [29‐31]. This evasive resistance implies adaptation of tumors to angiogenic inhibitors via activation of alternative pathways that sustain tumor growth.

Accordingly, our approach was designed to simultaneously target different angiogenic signaling pathways and to investigate the activation of possible resistance mechanisms in a GBM model.

In the present study, the antiangiogenic substances were delivered to a subcutaneous graft of G55 glioma cells using ex vivo modified PAE cells, which were encapsulated in alginate microbeads. The microencapsulation technology ensures a continuous release of proteins, and has been successfully used by us and others in different animal models [32, 33].

While these observations correlated with a pronounced decrease of vascular density in ES- treated tumors, treatment with Tum resulted in only minimal reduction of blood vessel density, suggesting that in vivo tumor growth reduction mediated by Tum is mainly caused by a direct antitumorigenic activities and less through antiangiogenic mechanisms. A direct αVβ3 -dependent growth-inhibitory effect of Tum on glioma cells in vitro and in vivo has been previously describe by Kawaguchi et al. [34]. On the other hand, the extent of tumor growth inhibition caused by the Es + Tum combination was higher than expected compared with the reduction level of vessel density. This fact prompted us to hypothesize that the ES + Tum combination exerts direct anti-neoplastic effects on glioma cells in vivo, in addition to its antiangiogenic effect. This hypothesis was confirmed in our in vitro experiments, which showed reduced proliferation rates of glioma cells after treatment with the ES + Tum combination, but not after treatment with the single inhibitors. Moreover, the ES + Tum combination caused morphological changes and induced apoptosis in glioma cells. Since previous studies have demonstrated that integrin antagonists affect cell cycle progression and viability of glioma cell lines, even inhibiting signaling pathways similar to ECs [34], we suggest that ES and Tum act through their respective integrin receptors on glioma cells, ultimately leading to inhibition of proliferation and induction of apoptosis. Nevertheless, further studies are necessary to clarify the effects of ES + Tum on glioma cells at the molecular level.

In order to gain further insights into possible mechanisms that enable tumor cells to escape anti-angiogenic therapies, we performed cDNA arrays using mRNA from tumor tissue treated with encapsulated PAE-WT cells or PAE cells releasing ES or Tum, either individually or in combination. Surprisingly, we identified only a few genes with a significant increase or decrease in expression level (mean signal log ratio ‡1.0) in the ES-, Tum- or ES + Tum-treated groups when compared with the control group. We focused our interest on the hormone prolactin (PRL) and its cognate receptor PRLR, which were up-regulated after treatment with Tum and ES + Tum, respectively. Validation of PRLR up-regulation in ES + Tum tissue sections by immunohistochemistry revealed a heterogeneous staining pattern with an intensive PRLR staining localized in well-defined tumor regions. Double immunostaining with apoptotic marker M30 and PRLR further showed that those areas with high levels of PRLR contained none or few apoptotic cells, whereas apoptotic regions presented low or no expression of PRLR. Similar results were obtained in vitro after immunofluorescence staining for cleaved caspase-3 and PRLR in glioma cells treated with ES + Tum (data not shown). Based on these results, we assume that a subpopulation within the G55 cells does not undergo apoptosis after ES + Tum-treatment but rather proliferates via activation of the PRLR/PRL signaling axis. In Glioblastoma “Cancer Stem Cells” (CSC), a small subpopulation of self-renewing “stem-like” cancer cells, have been demonstrated to show resistance to commonly used anticancer therapies such as radiation [35] and chemotherapy [36, 37]. Clark et al. [38] have observed a compensatory activation of multiple ERBB family receptors in GBM CSCs deprived of EGFR signal, suggesting an intrinsic GBM resistance mechanism for EGFR-targeted therapy. To what extent the PRLR-positive subpopulation found in ES + Tum-treated tumor consist of CSC needs to be further investigated in future studies.

Several studies have documented the involvement of the ligand PRL in the growth control of different tumors such as breast [39], liver [40] and prostate [41] and further, PRL antagonists such as hPRL-G129R has been demonstrated to inhibit breast cancer growth in vitro and in vivo [42]. However, only little is known about the role of the PRLR/PRL-signaling axis in glioma cells. PRLR expression has been found in rat and human glioma cells [28, 43] but also in benign intracranial tumors [44]. Ducret et al. [45] have shown that PRL induces a dose-dependent increase in proliferation and survival of U87-MG glioma cells. In line with these results we have detected PRLR mRNA expression in two additional glioma cell lines (G28 and G55) and could demonstrate that PRL stimulates cell proliferation in a dose-dependent manner, indicating that these cells express a functional PRLR.

Interestingly, we observed a strong up-regulation of PRLR in glioma cells treated with ES + Tum in vitro. PRLR expression in contrast was not influenced by oxygen deprivation as observed after incubation of G55 cells under hypoxic and normoxic conditions for 24 hours, 48 hours and 5 days (data not shown). These observations suggest that up-regulation of PRLR in GBM tumors after ES + Tum treatment was not a secondary response to the anti-angiogenic treatment, but rather mediated through direct action of both integrin targeting factors on tumor cells. Although little is known about the effects of ES and Tum on glioma cells at the molecular level, an integrin-mediated auto-regulation of cell proliferation and apoptosis in glioma cells have been recently described by our group and others [17, 34]. In addition, an integrin-PRLR cross-talk has recently been described in breast cancer cells [46]. The fact that cilengitide, an integrin αvβ3/αvβ5 inhibitor, partially blocked ES + Tum-mediated effect on PRLR expression point to an integrin dependent mechanism. It is therefore tempting to speculate that the combined application of ES and Tum triggers up-regulation of PRLR in glioma, resulting in augmented PRL signalling and ultimately in increased tumor growth and/or stimulation of angiogenesis [47]. Our in vitro data confirm to some extent this hypothesis as they show for the first time that PRLR overexpression significantly increases glioma cell growth. The PRLR-mediated increase of cell growth was abrogated by inhibition of Jak2, a tyrosine kinase that has been described as major downstream regulator of PRLR-signalling [28]. Moreover, we found a 4fold up-regulation of PRL expression in PRLR-overexpressing cells when compared to mock-transfected cells, suggesting a PRL-autocrine loop that stimulates glioma cell growth. Beside the already mentioned pro-proliferative activity of PRLR in diverse tumor entities, several groups have reported about a PRLR/PRL-mediated inhibition of apoptosis especially in response to chemotherapy. In breast cancer cells PRL confers resistance against cisplatin by activating a detoxification enzyme [48, 49] and in ovarian carcinoma cells PRL and its receptor inhibit apoptosis induced by serum starvation or cisplatin treatment [49]. These observations might explain the fact that ES + Tum-mediated cell growth inhibition in vitro was significantly less pronounced in PRLR overexpressing cells than in control cells.

Our current data demonstrate that the integrin inhibitors ES and Tum significantly reduce GBM growth in vivo. We also demonstrate that a simultaneous application of ES and Tum has more pronounced anti-tumorigenic effect than applications of each factor alone, and that this strong anti-tumorigenic effect of ES + Tum is likely mediated by a combination of anti-angiogenic and direct anti-tumorigenic activities. Moreover, we show that ES + Tum therapy induces up-regulation of the prolactin receptor in GBM in vivo and that the activation of PL/PRLR signaling stimulates proliferation. Additional studies are necessary to elucidate whether the PRL/PRLR signalling pathway represents a novel target for therapeutic strategies aimed at developing effective treatments for GBM.