Aiding and ABT’ing treatment for glioblastoma

Glioblastoma (grade IV astrocytoma) is the most common primary adult brain tumor. Although these tumors rarely metastasize, they almost always recur locally. In spite of intensive treatment regimens consisting of surgery, radiotherapy and temozolomide chemotherapy, patients with these tumors have poor prognoses with a median survival of under one year ^1,2. A number of factors may contribute to the resistance of these tumors to therapy. Conventional chemotherapy is limited by a relatively low drug penetration into the tumor interstitium, due to the need to cross the blood-brain and blood-tumor barriers ³. In addition, glioblastomas have regions of hypoxia ^4,5 that can lead to resistance to both chemo- and radiotherapy.

Despite the presence of hypoxic regions, one well-recognized hallmark of glioblastomas is endothelial cell proliferation and robust neovascularization (Figure 1). This is likely due to the expression of a variety of angiogenic growth factors such as vascular endothelial growth factor (VEGF). Recent evidence suggests that stem cell-like glioblasts may be a crucial source of key angiogenic factors and that targeting proangiogenic factors from these populations may be a viable therapeutic option ⁶.

Figure 1.

Hematoxyllin/eosin staining of human glioblastoma multiforme. Neoplastic cells stain blue and surround the area of central necrosis. Numerous microvessels are richly perfused with red blood cells.

Is the reliance of glioblastoma on angiogenesis the Achilles’ heel of this deadly neoplasm? Recently, this hypothesis has been put to rigorous tests in both pre-clinical and clinical settings. Xenograft models, where human glioma cells are implanted either ectopically (subcutaneously) or orthotopically (intracerebrally) into immunocompromised mice or rats, are particularly suited to assess the effects of anti-angiogenic molecules on tumor vessel density, overall growth, and survival ⁷. Numerous studies in animal models have shown that inhibiting VEGF function using neutralizing antibodies ⁸, dominant-negative VEGF receptor mutants ⁹ and antisense constructs ¹⁰ causes overt regression of blood vessels and thus precludes growth of glioma cells in vivo ¹¹.

Subsequently, a few promising studies have been performed in human patients. For example, AZD2171, an oral tyrosine kinase inhibitor of the VEGF receptor, has afforded significant clinical benefits in alleviating edema and normalizing the tumor vasculature ¹². However, in a recent Phase II clinical trial of bevacizumab (anti-VEGF antibody) plus irinotecan, six-month progression-free survival probability was only 38% ¹³. Thus, it appears that advanced glioblastomas do not rely exclusively on VEGF, and that other pathways involved in angiogenesis must be targeted in parallel.

One such approach would be the use of natural inhibitors of angiogenesis, whose action does not depend on neutralization of VEGF. One prominent member of this diverse family is thrombospondin-1 (Tsp1). This matricellular protein was the first naturally occurring angiogenic inhibitor discovered ¹⁴, and early on its anti-angiogenic effects were shown to limit tumor growth and metastasis ¹⁵. Many oncogenes, such as c-Myc, down-regulate Tsp1 to promote neovascularization ^16,17. Anti-angiogenic activity of Tsp1 is mediated primarily through its interaction with the scavenger receptor CD36 ¹⁸. This leads to the inhibition of endothelial cell migration ¹⁹ and the induction of p38 MAPK-dependent apoptosis ²⁰. At the same time, treatment with thrombospondin-1 results in increased expression of Fas ligand on the surface of endothelial cells, which makes them vulnerable to Fas-mediated apoptosis ²¹. Thrombospondin-1 may be a particularly useful therapeutic agent for glioblastoma because its expression is often absent in these tumors, due to frequent loss of chromosome 10 on which the THBS1 gene resides. Importantly, when chromosome 10 is re-introduced, human glioblastoma cell lines switch to non-angiogenic phenotype and thus lose their ability to form tumors in athymic mice ²².

The use of thrombospondin-1 as a cancer therapeutic has been limited by its very large size (>450 kDa as a trimer) and the presence of multiple functional domains ²³.

However, its anti-angiogenic potential is attributed mostly to the so-called type 1 repeats, or TSR ²⁴. In early pre-clinical studies, a recombinant protein encompassing all three TSRs was found to inhibit the growth of experimental B16F10 melanomas, Lewis lung carcinomas ²⁵ and human pancreatic cancer cells in an orthotopic mouse model ²⁶. Within TSRs, the anti-angiogenic activity has been mapped mainly to the DGGWSHWSPWSSC and GVITRIR amino acid sequences ²⁷, allowing the therapeutic use of even shorter peptides. The two modified peptides from the TSR region strongly suppressed tumor growth when administered intravenously to mice bearing MDA-MB435 breast carcinomas ²⁸.

In the last few years, Abbott Laboratories has developed and championed the use of another TSR-based therapeutic peptide, ABT-510, which is derived from the GVITRIR sequence ²⁹. Treatment with ABT-510 inhibits the growth of murine melanoma metastases in syngeneic animals and blocks the progression of orthotopic human bladder cell tumors ³⁰. Its potency is markedly improved when CD36 is upregulated using ligands of peroxisome proliferator-activated receptor gamma ³¹ and when combined with metronomic low-dose chemotherapy ³² or the histone deacetylase inhibitor valproic acid ³³. Last but not least, ABT-510 has proven safe in canine ³⁴ as well as human cancer patients ³⁵. Does it have the potential to become a drug of choice for glioblastoma?

In this issue of CB&T, Anderson et al describe the effects of ABT-510 on angiogenesis and tumor growth in mouse models of glioblastoma ³⁶. The authors observe that ABT-510 induces apoptosis of human brain microvascular endothelial cells (MvEc) by a CD36-dependent, caspase-8 mediated mechanism. It also inhibits tubulogenesis by MvEc propagated in 3D-collagen gels. More importantly, ABT-150 inhibits growth of human and murine glioblastoma grafts in orthotopic locations while decreasing microvascular densities. This work echoes an earlier study where a WSHWSPW-containing peptide was shown to significantly slow the growth of rat C6 glioma and 9L gliosarcomas ³⁷. While robust performance in mouse models certainly does not guarantee success in clinical settings, the authors’ data argue quite compellingly that ABT-510 should be evaluated in glioblastoma patients.