Evidences that zidovudine (AZT) could not be directly responsible for iron overload in AZT-treated patients: an in vitro study