Avoid using 1.1B4 cells they are a very poor beta-cell model. Other than that the ENDOCELLS that Amos refers to in his other answer are probably the best current available human beta-cell lines available for glucose-stimulated insulin secretion experiments. You will have to pay a lot of £££££ to get the MTA for these cells. Also worth note that they tend to grown very slowly like real beta-cells so plan ahead. The EndoBH-1 would be my preferred out of the 3 cell types ENDOCELLS offer. Note that you can also form pesudoislets out of these cells once they are growing well.