We consider how an integrated in vivo model can be used to study the specific transcriptional effects of specific receptors in neuroendocrine systems. Our example is the role of thyroid receptor (TR) isoforms in mediating negative feedback effects of T3 on TRH (thyrotropin releasing hormone) expression. The in vivo transfection method employed polyethylenimine (PEI) to introduce genes directly into specific regions of the brains of mice, rats, and Xenopus tadpoles. In the mouse model, the technique has served to examine TR effects on TRH transcription and on the pituitary-thyroid axis end point: thyroid hormone secretion. When a TRH-luciferase construct is introduced into the hypothalami of newborn mice TRH-luciferase transcription is regulated physiologically, being significantly increased in hypothyroidism and decreased in T3-treated animals. When various T3-binding forms of TRbeta or TRalpha are expressed in the hypothalamus, all TRbeta isoforms give T3-dependent regulation of TRH transcription, whereas TRalpha isoforms block T3-dependent transcription. Moreover, TR transcriptional effects are correlated with physiological consequences on circulating T4. Thus, somatic gene transfer shows TR subtypes to have distinct, physiologically relevant effects on TRH transcription. The approach is an appealing alternative to germinal transgenesis for studying specific neuroendocrine regulations at defined developmental stages in different species.