Brain metastases occur in approximately 15% of metastatic breast cancer patients and confer a dismal prognosis. Brain metastases are thought to increasing, particuarly among metastatic patients with Her-2+ or triple negative tumors. Our goals are to identify genes that functionally contribute to brain metastatic progression and to identify and validate preclinical leads. A murine preclinical model of brain metastasis was developed using a derivative of the MDA-MB-231 breast carcinoma cell line (231-BR cells). The proliferative and apoptotic rates exhibited by this cell line, as well as the presence of a neuro-inflammatory response, closely correlate with data from 16 resected brain metastases of human breast cancer, suggesting that the model is relevant. Additional models of brain metastasis of breast cancer have been estalblished to provide a measure of heterogeneity including the murine 4T1 mammary cell line (4T1-BR5) and Her-2 transfected human MCF-7 (ER+) cells (MCF-7-Her-2-BR3). In unpublished experiments, models of natural Her-2 overexpressing tumor cells, and inflammatory breast cancer cells are being developed for brain tropism. In collaboration with Drs. Quentin Smith and Paul Lockman, Texas Tech University, the permeability of experimental 231-BR and 4T1-BR5 metastases was quantified. Published collaborative research showed that experimental brain metastases of breast cancer were heterogeneous in their permeability, both within and between metastases in the same brain. While most metastases were permeable as compared to the normal brain blood-brain barrier, 10% exhibited sufficient permeability to enable a cytotoxic response to a systemic drug. The data argue that a new class of inherently brain-permeable drugs will be needed for this complication. We tested the hypothesis that Her-2 overexpression alters the natural history of breast cells to render them more brain metastatic. Her-2 transfectants of the 231-BR cells produced three fold greater large brain metastases, proportional to MRI detectable metastases in a human brain. The efficacy of the dual EGFR/Her-2 tyrosine kinase inhibitor, lapatinib, was tested on the brain metastatic colonization of a human breast carcinoma cell line.). Lapatinib prevented the brain colonization of 231-BR-Her-2 cells by 53% and inhibited the phospho-Her-2 staining of treated brain metastases in vivo, demonstrating that the drug hit its target. A pharmacokinetic analysis of lapatinib uptake in experimental brain metastases is submitted for publication. In these experiments, lapatinib resistant tumor cells were cultured from the brains of treated animals. The lapatinib resistant cells were sensitive to lapatinib inhibition of proliferation in vitro, suggesting that lapatinib resistance in the brain is mostly due to inadequate drug delivery. Rational combinations with lapatinib were investigated to improve its preventive efficacy. Pazopanib, a VEGFR, PDGFR and c-kit inhibitor, was investigated. Pazopanib was anti-proliferative to 231-BR-Her-2 cells as well as endothelial cells. Inhibition of pErk was observed in the tumor cells despite the fact that they harbored mutations in both Ras and B-Raf. Enzymatic assays demonstrated that pazopanib has B-Raf inhibitory activity distinct from previously reported B-Raf inhibitors. It is preferentially inhibitory to Her-2 activated wild type B-Raf, as compared to the common V600E melanoma mutation. In vivo pazopanaib significantly prevented the brain metastatic colonization of both 231-BR-Her-2 and MCF-7-Her-2-BR3 cells. Analysis of the treated brain metastases showed a dimunition of pErk and PMek tumor cell staining, with no dimunition of blood vessel density, suggesting that the B-Raf inhibitory activity of pazopanib was functionally important. Lapatinib/pazopanib combination experiments are under consideration. Affiliated work on pazopanib, examining primary breast and melanoma tumors, is submitted for publication. This work demonstrates that the B-Raf inhibitory activity of pazopanib is also correlated with its anti-angiogenic activity. In collaboration with Dr. George Sledge, Indiana University, a 13-gene signature was developed from gene expression profiling of primary tumor material from metastatic Her-2+ breast cancer patients, that predicts short (3 years) time to the development of brain metastases. This signature has now been independently validated and is being prepared for publication. Included among the 13 genes were the overexpression of three DNA double strand break (DSB) genes, Rad51, Bard1 and FancG, all overexpressed in tumors rapidly developing brain metastases. The role of Rad51 and Bard1 overexpression is under investigation in the 231-BR model. To date, our preliminary data indicate an exciting phenotype in the absence of DNA damage inducing drugs. A microarray analysis of surgically resected brain metastases of breast cancer, using laser capture microdissection, amplification and 30K cDNA arrays. These data were compared to a cohort of unmatched primary breast tumors, matched for histopathology, TNM and grade. A heat map comparing gene expression differences between brain metastses and unmatched primary tumors has been compiled and expression trends validated by QRT-PCR using an independent cohort. Of the genes validated, experiments are ongoing for hexokinase 2 (HK2 and pigment epithelium derived factor (PEDF). PEDF is a secreted protein with anti-angiogenic, tumor suppressor and neuronal viability properties, possibly mediated through multiple receptors. Overexpression of PEDF in the 231-BR cell line reduced the number of large parenchymal metastases significantly. Due to the instability of the PEDF transfection construct in vivo, intracranial injections were used for analysis of an interaction with the microenvironment. PEDF overexpressing tumor cells formed smaller intracranial tumors then control transfectants. Interestingly, neuronal damage, assessed by two independent neuropathologists, was reduced in the zone surrounding the tumor cells overexpressing PEDF. Neuronal damage was assessed by traditional staining and a new Flourojade technology that the lab has adapted to cancer studies. PEDF therefore stands as a potential strategy to reduce the damage to surrounding nerves caused by developing metastases. This work is under revision for publication. In preclinical experiments, 16 potential therapeutics have been tested for prevention of 231-BR brain metastatic colonization. Of these, only four have demonstrated efficacy, highlighting the prohibitive role of the blood-brain barrier. Where tested, the active drugs prevented brain metastases, but statistically failed to treat (shrink) an established brain metastasis. Considering that all clinical trials are conducted with the primary endpoint of responses in established lesions, the data indicate that we are missing some potent activities with preventive potential. New clinical trial designs are needed.
Showing the most recent 10 out of 38 publications
