there is a method to measure the antenna gain called "Purcell method", which involves a single antenna looking at a reflecting plane in front of it. From S11 measurements you can extract a line and the slope is proportinal to the gain.
If you have access to IEEExplore look for that. There are several papers on it , for instance,
Z. Wu, "Effect of Mismatch on Antenna Gain Measurement by Purcell's Method," Electronics Letters, vol. 22, pp. 522-524, 1986.
I dont know any method measuring S21 between two antennas, although probably you may think of something varying the distance between antennas.
regards