For the same reason then why not "unique key". Also, I don't know in theory unique key is there in SQL.
The whole restriction of foreign key is that the values in child table can come only from the parent table. Like manager being an employee in an office implies, managerid must come from set of employeeid values. When a manager employee resigns, the manager table entry also becomes invalid. So, it should be A, B and C and "only" should make C the answer.
Even D is possible as in MYSQL, but the issue is when parent table entry is having duplicates, when an entry is deleted from parent table, we cannot simply delete the same (delete on cascade) from child table.