Remarks on a nonhomogeneous model of magnetohydrodynamics