They could at least send the paper with the reported problems out to a new set of referees.

And just as they decided to take responsibility for publishing, they can take responsibility after a similar review for retraction (or issuing an errata or whatever fancy way they want to signal the result of the process).