Not that I can think of. Consider one of those columns has an index. If you update syscolumns with a new collation (say by going from case sensitive to case insensitive), that index is now "corrupt". And that does not even get into the idea of unique indexes, which could become non-unique after such a change.
As painfull as it sounds, you will have to go through and do all of the thousands of columns, and rest their collations with some sort of script.
If I might ask, how did this requirement come about?