Density functional theory (DFT) studies on two polynuclear clusters, [CuII 5GdIII 4O2(OMe)4(teaH)4(O2CC(CH3)3)2(NO3)4] (1) and [Cu5Gd2(OH)4(Br)2-(H2L)2(H3L)2(NO3)2(OH2)4] (2), have been carried out to probe the origin of the large magnetocaloric effect (MCE). The magnetic exchange interactions for 1 and 2 via multiple pathways are estimated using DFT calculations. While the calculated exchange parameters deviate from previous experimental estimates obtained by fitting the magnetic data, the DFT parameter set is found to offer a striking match to the magnetic data for both complexes, highlighting the problem of overparameterization. Magnetostructural correlations for Cu-Gd pairs have been developed where both the Cu-O-Gd angles and Cu-O-Gd-O dihedral angles are found to significantly influence the magnitude and sign of the exchange constants. The magnitude of the MCE has been examined as a function of the exchange interactions, and clues on how the effect can be enhanced are discussed.