Why it is true that, over an algebraically closed field, any abelian variety is isogenous to a principally polarized abelian variety?