Brain metastases of breast cancer are thought to increasing in incidence, particuarly among metastatic patients with Her-2+ or triple negative tumors, and confer a dismal prognosis. Our goals are to identify pathways that functionally contribute to brain metastatic progression and to identify and validate preclinical leads. I. Five experimenal brain metastasis assays were developed or co-developed in the lab: Human MDA-MB-231BR triple negative tumor cells; murine 4T1-BR mammary cell line; Her-2 transfected human MCF-7 (ER+) cells (MCF-7-Her-2-BR3), and naturally Her-2+ human Jimt-1-BR and SUM190-BR cells. II. 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 blood-brain barrier, only 10% exhibited sufficient permeability to enable a cytotoxic response to a systemic drug. We have completed a quanitative immunofluorescence study of the blood-tumor barrier (BTB) in three model systems. The relative expression of blood-brain barrier components were compared between normal (uninvolved) brain and brain metastases, and also between brain metastases that were either poorly or highly permeable to Texas red dextran. Many alterations were observed between normal brain and the BTB, including endothelial cell size, neuro-inflammation, absence of astrycyte endfeet, increased VEGF, etc. In contrast, few differences were observed between poorly and highly permeable metastases. The relative expression of subpopulations of pericytes was the major difference observed. Highly permeable lesions exhibited an increase in Desmin+ pericytes and a decrease in CD13+ pericytes. Experiments will not determine if these alterations functionally contribute to altered permeability. III. Permeable and impermeable brain lesions have also been identified in frozen sections of mouse brains; using the adjacent section the material was laser-capture microdissected. RNA extracted from these lesions was hybridized to both human (tumor cell genes) and mouse (brain microenvironment genes) microarrays. The goal of these experiments are to identify differentially expressed genes between permeable and impermeable lesions.All are from the microenvironment, rather than the tumor cells. This project is expected to provide an agnostic analysis of permeability in brain metastases. It is hoped that these analyses will lead to our ability to better permeabilize these lesions to therapy. IV. We published experimental brain metastasis data indicating that temozolomide, a first line treatment for primary brain tumors, was 100% effective at preventing brain metastases of 231-BR cells over a two log dose response. This type of profound prevention has not been previously observed. Given later when brain metastases are already at least partially established, temozolomide was ineffective. Activity was dependent on methylguanine methyltransferase (MGMT). Staining of matched sets of primary breast cancers and resected brain metastases showed poor concordance, but 60% of the brain metastases were low in MGMT. The lab has now demonstrated brain metastasis preventive effects of lapatinib, pazopanib and temozolomide for brain metastasis of breast cancer in preclinical models. A phase II trial is being developed in the WMB to test these interventions. V. Comparison of matched human primary tumors and resected brain metastases revealed that Rad51 and another DNA double strand break repair protein, Bard1, were overexpressed in the brain lesions. We have found that transfection of either gene into 231-BR or 4T1-BR cells increases brain metastsis formation by approximately four fold. No effect on lung metastasis was found. The augmentation of brain metastasis by either Rad51 or Bard1 overexpression was found to be dependent on an oxidative response in the brain microenvironment, likely causing DNA damage. Tempol, an oxygen radical scavenger, abrogated the effects of Bard1 and Rad51 overexpression. Current experiments will determine the effect of Bard1 or Rad51 overexpression on radiation therapy and chemoprevention.
Showing the most recent 10 out of 38 publications
