The core question is
kodama said:
what special properties does a massless spin-2 boson cause time dilation that a massless spin-1 boson does not?
If I consult the Feynman Lectures on Gravitation, as suggested in the link I posted, I find Feynman contrasting particle motion in an electromagnetic field and in a gravitational field, as follows: "the gravitational equation has a qualitatively distinct new feature; not only the gradients, but also the potentials themselves appear in the equations of motion" (5.2.1).
From this he derives the gravitational time dilation. It seems to be a "timelike" manifestation of the fact that gravity is sensitive, not just to potential energy differences ("the gradients"), but to the absolute amount of energy ("the potentials themselves").
Meanwhile, I believe this universality of gravity can be deduced quantum mechanically, by considering the amplitude for emitting a massless spin-2 particle in the limit where the particle has zero momentum, we show that the coupling to all forms of energy-momentum is the same (I am referring to "Weinberg's low-energy theorem"). The analogous argument for a massless spin-1 boson only implies charge conservation.
So, my schema for explaining this is, low-energy theorem implies graviton coupling is universal, which implies a kind of absolute sensitivity to energy-momentum, and in particular that the speed of a physical process depends on the gravitational potential.
I am not 100% sure that I have it right, even schematically; and there is probably also a more geometric account of this, based on spin-2 mapping to a metric, but spin-1 only to a connection. But the combination of Weinberg and Feynman, correctly interpreted, surely has the kernel of an answer.