Some remarks on Brauer groups of K3 surfaces