2023A&A...679L...9V

Recent studies of high-redshift galaxies with James Webb Space Telescope (JWST), such as GN-z11 at z = 10.6, show unexpectedly significant amounts of nitrogen (N) in their spectra. As this phenomenology appears to extend to gravitionally lensed galaxies at Cosmic noon such as the Sunburst Arc at z = 2.37, as well as globular clusters overall, we suggest that the common ingredient among them are very massive stars (VMSs) with zero-age main sequence (ZAMS) masses in the range of 100-1000 M_⊙. The He II in the Sunburst Arc might also be the result of the disproportionally large contribution of VMS to the total stellar contribution. We analyse the pros and cons of the previous suggestions, including classical Wolf-Rayet (cWR) stars and supermassive stars (SMSs), to conclude that only our VMS alternative ticks all the relevant boxes. We discuss the VMS mass-loss history via their peculiar vertical evolution in the HR diagram resulting from a self-regulatory effect of these wind-dominated VMSs and we estimate that the large amounts of N present in star-forming galaxies may indeed result from VMSs. We conclude that VMSs should be included in population synthesis and chemical evolution models. Moreover, that it is critical for this to be done self-consistently, as a small error in their mass-loss rates would have dramatic consequences for their stellar evolution, as well as their ionising and chemical feedback.