This has been commonly observed. Mutated forms of tumor suppressors can behave as tumour promoters. For example some TP53 mutants act as inhibitors of wild-type TP53. Read for example Tumor Suppressors Having Oncogenic Functions: The Double Agents. Also tumor suppressor genes could be expressed in response to normal signals but be inactive due to change of context (e.g. proliferating tumor cell vs quiescent wild-type) or mutations (in them or in downstream effectors). Another possibility is that the stability of the protein has been affected so levels are higher than in wild-type cells. I am sure I am forgetting other explanations but the literature should be full of examples.