@LYg: Yes, $q_{\mu}\Pi^{\mu\gamma}(q)=0$ is a result of local gauge symmetry. It can be derived from equation 7.107 in the first edition of Ryder (may be different in your edition). This is equation relates a derivative of a gauge field to functional derivatives of $\Gamma$ with respect to respect to the gauge field and with respect to fermion fields. If you functionally differentiate this equation with respect to the gauge field and then set the fermion field to zero, you obtain a relation between a gauge term and a derivative of the inverse photon propagator which gives the desired result.