What limits the size of an accelerator is usually the cost. In the case of the SSC, the size was determined by the physics requirements. In the case of the LHC, the limit was the construction cost. In the case of LEP, the lower bound limit to the size was the cost of powering the thing.
As you accelerate charged particles, they emit photons. Particles with a large charge-to-mass ratio, such as an electron, will emit photons with gamma-ray energies. This synchrotron radiation is constantly being shed by the beam, and that energy has to be replaced by the accelerator. The power radiated goes as E5/r2, where E is the beam energy and r is the radius of the accelerator. You can see that as the size gets larger, the cost of powering it actually goes down. (As the beam energy goes up, however, the cost goes dramatically up.)
The particles move so fast that gravity is a very low order factor, essentially zero. However, gravity has an effect on the apparatus itself. Everything has to be aligned to a gnat's whisker and gravity is constantly putting parts under tension and compression, so, no doubt, things move and have to be checked all the time and readjusted. In space there wouldn't be nearly the same forces aside from natural resonances, springiness, and thermal effects due to heating and cooling. Alignment will probably be a major problem in space, too.