internally-headed relatives with external heads

[Yury Lander]

Dear typologists,

1. Does anyone know any internally-headed relative construction where the internal head is marked with the external case?
Such a construction was once found in Shapsug Adyghe (I mentioned it in my paper in Rice WPL), but even there it seems to be a "mix" of an externally-headed construction and an internally-headed construction. So I wonder whether it is indeed an outstanding story.

2. And yet another question. In Adyghe, the internally-headed construction structurally allows an external head (i.e. in theory, a relative may have two semantic heads, lexically different, although many speakers consider this a too complex structure). Please let me know if you are aware of similar cases.

Thanks in advance,
Yura