Coordinated programs of gene expression during cell differentiation can be controlled by master transcription factors. The differentiation of helper T (Th) lymphocytes during the immune response has been shown to occur along alternative pathways designated as Th1 and Th2. Induction of the Th1 and Th2 pathways is associated with the conversely regulated expression of the master factors T-bet and GATA-3, respectively. Both autoactivation and inhibition of GATA-3 play a crucial role in this process. We develop mathematical models of the underlying regulatory networks to provide a framework for the analysis of experimental data. Modeling concepts for gene expression dynamics are introduced, and paradigms for the behavior of gene-regulatory networks are reviewed. A mechanistic model for the regulation of GATA-3 in Th cells is developed that accounts for autoactivation and regulation by external differentiation signals. This system works as a bistable switch that enables the triggering of a differentiation program by transient inductive signals. GATA-3 inhibitors (such as FOG-1 and ROG) modulate GATA-3 expression by yet unidentified mechanisms. Three potential modes of inhibition, sequestration by a binding protein, repression of basal transcription, and repression of autoactivation, are predicted to have distinct, and strongly concentration-dependent, regulatory effects on GATA-3 dynamics. Based on these results, we develop a model for the cross-regulation of the alternative Th1 and Th2 differentiation programs which are governed by the dynamics of T-bet and GATA-3, respectively. The steady states of this model correlate with naïve, Th1-polarized, and Th2-polarized phenotypes. Our analysis makes predictions on the stability of the Th1 and Th2 programs and raises questions on the relation between transcription factor regulation and epigenetic determination in cell differentiation.