UPDATE
As I predicted, the pin is not connected internally as confirmed by Clare/IXYS:
The pin is floating so there is no internal connection.
Best regards,
Anthony Konopka
IXYS Integrated Circuits Division
78 Cherry Hill Drive
Beverly, MA 01915
Tel: (978) 524-6742
Fax: (978) 524-4700
E-mail: akonopka@clare.com
PREVIOUS
It's difficult to say. The datasheet doesn't offer any insight. I wrote to tech support. I'll update my answer if they respond.
The pin may be:
used for functional testing at the production facility and therefore connected internally to something that might interfere with your signal
it might be left unconnected (floating)
it might be shorted to another one of the pins (usually some type of reference like "ground")
If you can accept destructive modification to the part, just clip the pin as suggested by @MichaelKohne. If you can't, you are going to have to mitigate the risk. This will involve doing things that are outside best practices for mass production and high-reliability applications, however, they may show that you can do what you wanted in the case of this one unit in your home project for your personal use (where you can competently accept the risk):
test for cross-pin connectivity with an Ohm-meter
short the pin through an ammeter and look for current
apply voltage to the pin and see if there is current flow
power up the relay and operate normally, monitor the pin with an oscillioscope to look for any drive behaviors.
If I had to guess, I would say that it is probably fine. I'm not saying to do it blindy. I'm saying test it... and if I had to speculate as to what your tests will show, I would guess that they show it's ok to connect. I will clarify the reasoning:
You imply by your question that you want to use the pin (so let's see if we can make that happen for you!). If that isn't your intention then this is a silly question. Do not use a "do not use" pin doesn't require asking on SE, so I consider those responses (the ones that don't include reasoning other than "the manufacturer says so") to be irrelevant.
DNU can mean NC (and often does). DNU can be synonymous with "reserved for future". Since future part revisions are not a concern on a 1-off project, it's ok in my judgement if it clears the electrical test hurdles.
There are no countermanding standards that I am aware of. The meaning and application of these terms are manufacturer/part specific. If you look at similar parts from this manufacturer the trend is to cite "do not use" whenever the number of functional pins is exceeded by package pin count. Having designed many dies, I believe that it is unlikely that each of these parts has a hidden test access mode requiring a unique dedicated pin.
BUT... I may be wrong. It's my decision calculus. If you have a different one. That's fine. That is why I was careful to identify that this last part is just my speculation and why I contacted the manufacturer for clarity.