3047 Background: Antibody (AB) therapy has been successful against breast cancer, but brain mets are not well controlled. One likely factor is that AB molecules are too large to cross the blood-brain barrier (BBB). One approach is to truncate the AB, but this can affect its function. We propose an alternative. Unlike inert proteins, cells can cross the BBB. Indeed, antibody-secreting cells (ASC) are found in the brain in many neurologic disorders. Our hypothesis is that ASC, lodged within the brain, can be a source of AB to attack both pre-existing and new brain mets. We describe a new rat model to test this, and initial findings. Methods: Rat mammary carcinoma 13762 was made to constitutively express surrogate tumor antigens, either b-galactosidase (b-gal) or alkaline phosphatase (ap). Tumor cells were injected into the left common carotid artery of syngeneic CDF rats, to favor delivery to the brain. Tumor is identified by staining for keratin or the surrogate antigen. In a new assay, ASC are identified by a histochemical stain for their cognate antigen. As a positive control for ASC staining, spleens of immunized rats are used. Results: 1) The new histochemical assay does identify ASC correctly in positive and negative control spleens. 2) In the brains of rats that had been pre-immunized to b-gal, ASC specific for b-gal were seen at sites of b-gal+ tumor. 3) Some ASC were also seen away from tumor. Conclusions: Although ASC are known to enter the brain in many neurologic disorders, little work has stressed brain metastases. This new model allows study of ASC traffic to the brain in a tumor context. Favored sites of ASC entry and retention can be defined in both tumor-free brains (as appropriate for prophylaxis) and brains with pre-existing metastases. No significant financial relationships to disclose.