There’s no automated way of doing this in the engine at the moment that I know of.
So, I would create 3 versions of the audio file. One clean version, 2nd, a distorted version, and 3rd, white noise.
Then you could use a crossfade by distance node in the sound cue to fade between all 3.
It would be pretty easy to create the distorted version. Apply EQ to cut high and low frequencies, amplify, then use a limiter with hard clip, apply a high pass filter, amplify again and limit etc. There’s a lot of tutorials on how to achieve such an effect.
Then generate white noise in some audio program like Audacity.
Import all the waves, and set up a sound cue like the below:
You’ll have to play around with the distances until you get the effect you want. The attenuation node allows the sound to be spatialized and attenuated. Adjust that as needed also.
Example of what the effect could sound like:
https://drive.google.com/open?id=0B3KowC8HYfyeUGJadXBlaFRDU3M