We investigated the presence of EphA2, and its ligand, ephrinA1, in glioblastoma multiforme (GBM), a malignant neoplasm of glial cells, and normal brain. We also initially examined the functional importance of the interaction between EphA2 and ephrinA1 in glioma cells. Expression and localization of EphA2 and ephrinA1 in human GBM and normal brain were examined using Western blotting, immunofluorescence, and immunohistochemistry. A functional role for EphA2 was investigated by assessing the activation status of the receptor and the effect of ephrinA1 on the anchorage-independent growth and invasiveness of GBM cells. We found EphA2 to be elevated in f90% of GBM specimens and cell lines but not in normal brain, whereas ephrinA1 was present at consistently low levels in both GBM and normal brain. EphA2 was activated and phosphorylated by ephrinA1 in GBM cells. Furthermore, ephrinA1 induced a prominent, dose-dependent inhibitory effect on the anchorage-independent growth and invasiveness of GBM cells highly overexpressing EphA2, which was not seen in cells expressing low levels of the receptor. Thus, EphA2 is both specifically overexpressed in GBM and expressed differentially with respect to its ligand, ephrinA1, which may reflect on the oncogenic processes of malignant glioma cells. EphA2 seems to be functionally important in GBM cells and thus may play an important role in GBM pathogenesis. Hence, EphA2 represents a new marker and novel target for the development of molecular therapeutics against GBM. (Mol Cancer Res 2005;3(10):541 -51)

This report describes the methodology, diagnostic yield, and adverse events (AE) associated with frame-based stereotactic brain biopsies (FBSB) obtained from 26 dogs with solitary forebrain lesions. Medical records were reviewed from dogs that underwent FBSB using two stereotactic headframes designed for use in small animals and compatible with computed tomographic (CT) and magnetic resonance (MR) imaging. Stereotactic plans were generated from MR and CT images using commercial software, and FBSB performed both with (14/26) and without intraoperative image guidance. Records were reviewed for diagnostic yield, defined as the proportion of biopsies producing a specific neuropathological diagnosis, AE associated with FBSB, and risk factors for the development of AE. Postprocedural AE were evaluated in 19/26 dogs that did not proceed to a therapeutic intervention immediately following biopsy. Biopsy targets included intra-axial telencephalic masses (24/26), one intra-axial diencephalic mass, and one extra-axial parasellar mass. The median target volume was 1.99 cm3. No differences in patient, lesion, or outcome variables were observed between the two headframe systems used or between FBSB performed with or without intraoperative CT guidance. The diagnostic yield of FBSB was 94.6%. Needle placement error was a significant risk factor associated with procurement of non-diagnostic biopsy specimens. Gliomas were diagnosed in 24/26 dogs, and meningioma and granulomatous meningoencephalitis in 1 dog each. AE directly related to FBSB were observed in a total of 7/26 (27%) of dogs. Biopsy-associated clinical morbidity, manifesting as seizures and transient neurological deterioration, occurred in 3/19 (16%) of dogs. The case fatality rate was 5.2% (1/19 dogs), with death attributable to intracranial hemorrhage. FBSB using the described apparatus was relatively safe and effective at providing neuropathological diagnoses in dogs with focal forebrain lesions.

The vast majority of brain cancers (gliomas) express a receptor (R) for interleukin 13 (IL13). In order to achieve specific targeting of the IL13R in gliomas, we have mutagenized human (h) IL13. The mutation was made to alter IL13 interaction with the shared functional IL13/4 normal tissue receptor, but not with the glioma-associated receptor. We have thus produced hIL13.E13K (glutamic acid at position 13 changed to lysine) and fused it to derivatives of Pseudomonas exotoxin A. The hIL13.E13K-based cytotoxins are less active on normal cells and thus less toxic, and are better antitumor agents compared with the cytotoxins containing nonmutagenized hIL13.

We have found that binding sites for interleukin-13 (IL-13) are overexpressed in a vast majority of high-grade astrocytomas (HGAs). These binding sites for IL-13 are distinct from the physiological receptor in that it does not bind IL-4. We also demonstrated that IL-13 receptor alpha 2 protein chain (IL-13Ralpha2), an IL-4-independent receptor for IL-13, is abundant among HGAs, but not in normal organs. To examine if IL-13Ralpha2 is the tumor-associated site for IL-13, we stably transfected normal Chinese hamster ovary (CHO) cells and glioma G-26 cells to express either human (h) or murine (m) IL-13Ralpha2. CHO-hIL-13Ralpha2(+) cells and G-26-h/mIL-13Ralpha2(+) cells, and not CHO and G-26 parental or mock-transfected cells, specifically bound IL-13 in an IL-4-independent manner. The IL-13Ralpha2(+) cells also became highly susceptible to the killing by an IL-13-based cytotoxic fusion protein. In loss of function studies, a HGA cell line, SNB-19, was transfected with antisense (as) hIL-13Ralpha2. as-SNB-19-hIL-13Ralpha2(+) cells lost their natural affinity towards IL-13 and became resistant to IL-13-based cytotoxins. The fact, that IL-13Ralpha2-positive cells bind IL-13 independent of IL-4, become susceptible to IL-13 cytotoxins, and cells deprived of IL-13Ralpha2 receptor lose these features, demonstrates that IL-13Ralpha2 is the brain tumor-associated receptor for IL-13.