Laser capture microdissection (LCM) holds great potential for analyzing gene expression profiles in situ. Most recently, this laboratory employed a novel immunostain-based LCM protocol (immuno-LCM) to selectively retrieve brain microvascular endothelial cells (BMEC) from intimately associated perivascular cells. However, before this protocol can be confidently coupled to downstream analytical platforms, it must be demonstrated that any variability associated with it is minimal, so as not to obscure data interpretation. As various factors could contribute to variability, this study focused on determining whether technical inconsistency and/or biological diversity of sample populations, played such a role. Specifically, two separate immuno-LCM-derived BMEC samples derived from adjacent tissue sections of a single mouse (to detect only technical variability), and from analogous tissue sections of three different mice (to detect technical and biological variability) were compared for their relative expression of 16 genes, using quantitative-RT-PCR (qRT-PCR). Both significant linear and rank-order correlations were observed between different sections from the same animal, underscoring lack of technical variability in this LCM application. Furthermore, a three-dimensional scatter plot of gene expression profiles from the three animals was linear, and ANOVA showed absence of statistically significant differences between any of the animals, confirming lack of biological variability. These findings argue that immuno-LCM coupled to qRT-PCR affords a reproducible means to assay gene expression in situ.