Remarks on the gauge-invariant formulation for interaction of radiation with matter