Could this still work?
I´m applying the pressure in the right direction, but AIR pressure (less density than water...)
If you're designing something to work with o-rings, I've found this handbook to be invaluable:
http://www.parker.com/literature/ORD%20570...ng_Handbook.pdf
If you follow the recommendations in that handbook, your seals will work, they will work well, and they will work for a long time. The handbook is conservative, in that you can break many of their rules and still get a good reliable seal, but it will require more testing to prove that. For example, they recommend against using silicone rubber with saltwater, but there are many reliable underwater camera housings that do exactly that.
Getting back to your question, if you're testing in anything other than real-world, worst-case conditions, you need to decide what type of failure you're testing for. In my mind, these are the main causes of seal failure, in no particular order:
1. contamination
2. bad surface finish or hairline scratches
3. seal extrusion
4. the seal not seating properly in the groove
5. worn out seals
6. the housing bending so that the seal doesn't fit into the groove
7. chemical attack on the housing, shaft, or seal
Your test with air will actually be better than testing with water for something like #2, the hairline cracks. Because water beads up on rubber seals, water doesn't leak through some of the smallest cracks, but air does (this is the principle used in waterproof/breathable fabrics like Gore-Tex). So testing with air and the buttons backwards will be a conservative test as far as detecting surface scratches or cracked seals.
On the other hand, you also want to test the actual surfaces that will be used, and you want to make sure that the housing bends in the way you expect it to bend, and you want to make sure that small amounts of contamination or silt or crystallized salt won't compromise your seal.