Exploring Iterative Enhancement for Improving Learnersourced Multiple-Choice Question Explanations with Large Language Models