@cjquijano - I feel the results you have got are logically very correct. When you add prism layers, effectively , the boundary layer velocity calculation is made more accurate. The cell relative velocity decreases and hence the predicted drag is higher. Parallely, the lowered cell relative velocity affects the values in some cells above, hence lower lift value computation.
But yes,as @Milad_Mafi Milad_Mafi stated, it is very important to take note of the y plus values as most turbulence models have a numerical domain of y plus in which they work prefectly.