Original video assets from a studio has multiple audio tracks, but they are transcoded down to a single audio track before distribution to the end user. Sorry.
You could do some fancy editing, or maybe you can use machine learning somehow.
But removing the laugh track really messes with the comedic timing.
BUT BUT BUT THOSE ARE ILLEGAL TO SHARE