Long-term in vitro growth of human T cell clones: can postmitotic 'senescent' cell populations be defined?