NLPGuard: A Framework for Mitigating the Use of Protected Attributes by NLP Classifiers