DELL: Generating Reactions and Explanations for LLM-Based Misinformation Detection