> it seems that the interstage transformer could be replaced with R-C coupling
With R-C coupling, as it starts to limit, there is common-mode shift that the second stage has to ignore. If you kept the 175's 5K plate resistors, it idles at about 6mA per side but that falls to about 1.5mA to get 6dB GR. 6-1.5=4.5mA, 4.5mA*5K= 22V of Common Mode shift. Hard to get that much without a current source in the next stage. The transformer eliminates the issue.
BTW, there are many R-C coupled limiters and some of them seem to deliberately clip in the second stage grids as they go into GR. If you limit only as emergency protection for a transmitter, transient clipping is easier than a faster attack (higher power rectifier amp). This can also be good "flavor" when trying to boost your average/peak ratio in a dense mix.