Delete pandas rows checking if values is in other all other column values series

Question

If I have a dataframe like this:

index   col1        col2
0       madrid      barcelona
1       barcelona   gracia
2       valencia    puerto
3       madrid      valencia
4       madrid      centro

Is it possible to delete all rows which have a col2 value that is in col1 values (checking all values of the col1, not only its row)? In the example it would delete rows 0 and 3.

I'm trying 'isin' function but it is not working.

Thank you.