Long gamma-ray bursts are supposed to be connected to the death of very massive stars. Due to their brightness, we can detect them to much larger distances than supernovae. Using them as powerful lightsources, they allow us to study star-forming high redshift galaxies and their interstellar medium in great detail with medium and high resolution spectroscopy. Despite the large redshift ranged spanned by GRBs, there is surprisingly little evolution in the properties of their host galaxies which might indicate that GRBs can only occur under certain conditions. This can be investigated from a few bursts at very low redshifts where we can resolve their host galaxies e.g. with integral field spectroscopy. The immediate surroundings might allow us some conclusions on the progenitors of GRBs.