Microsatellite instability is responsible for the human repeat expansion diseases (REDs). The mutagenic process differs from classical cancer-associated microsatellite instability (MSI) in that it requires the mismatch repair proteins that normally protect against MSI. LIG4, an enzyme essential for non-homologous end-joining (NHEJ), the major pathway for double-strand break repair (DSBR) in mammalian cells, protects against expansion in mouse models. Thus, NHEJ may compete with the expansion pathway for access to a common intermediate. This raises the possibility that expansion involves an NHEJ-independent form of DSBR. Pol θ, a polymerase involved in the theta-mediated end joining (TMEJ) DSBR pathway, has been proposed to play a role in repeat expansion. Here we examine the effect of the loss of Pol θ on expansion in FXD mouse embryonic stem cells (mESCs), along with the effects of mutations in Rad52, Rad54l and Rad54b, genes important for multiple DSBR pathways. None of these mutations significantly affected repeat expansion. These observations put major constraints on what pathways are likely to drive expansion. Together with our previous demonstration of the protective effect of nucleases like EXO1 and FAN1, and the importance of Pol β, they suggest a plausible model for late steps in the expansion process.