I think the original problem can be solved using the identity
which holds for all positive integers m,n. This identity can be proved by fixing m, say, and inducting on n. Then, prove that
divides
with another inductive argument, using the above identity.