Massive galaxy clusters are important probes of structure formation and cosmology, and are also natural gravitational telescopes which can bring lensed primordial galaxies from the earliest epochs into confortable reach for detailed studies. The CLASH project (Cluster Lensing And Supernova survey with Hubble) combines an HST Treasury program to obtain panchromatic imaging of 25 massive clusters, with other multi-wavelength observations, primarily spectroscopy with a Large Programme with VLT/VIMOS. This program was designed to provide landmark progress in the investigation of the dark matter mass distribution of galaxy clusters and accurate mass density profiles on the Kpc-Mpc scale, thus testing distinctive predictions of the LCDM cosmological model. After showing the status of the CLASH-VLT project and a number of relevant results recently obtained in this area, I will focus on the use of CLASH clusters as gravitational telescopes to detect the most distant galaxies at the edge of the reionization epoch. I will show how this strategy is proving to be an efficient way of unveiling galaxies out to z~11 to date, providing the first estimate of the star-formation rate density some 500 Myrs after the big bang. These first results have stimulated a new HST observational campaign, the Frontier Fields, which will further expand our knowledge on the formation of the first energy sources in the Universe.