result = cellfun( @(x) x{1}, cellfun( @(x) strrep( strsplit(x), ',', '' ), myCell, 'UniformOutput', false ), 'UniformOutput', false );
works I think, but it doesn't win any prizes for neatness. You can split it over multiple lines to be a bit less daunting if you prefer.
No need for regexp though.
If your values are always just 2 letters then you can do simpler things, just accessing the first 2 elements of the cell and ignoring what comes next, but obviously that fails in the generic case.
