There's already a demonstration of this with HIV. Most HIV proteins seem to have a bit of room to mutate and thus no longer get presented on MHC, ie there are 'escape mutants.' But there's one particular part of one particular protein that isn't so tolerant of mutations. People whose MHC can present that peptide mount a much more effective response against the infection.
Correct.
Yes, there are (broadly) two types of MHC.
The kind of MHC in this case is found on almost all nucleated cells. What's supposed to happen is that any/all of the proteins made in a cell get broken down into peptides at some rate and some of these peptides get presented on the MHC. It's kind of an ID card for cells. If the cell is infected with a virus, then viral peptides can make their way into the MHC. Killer T cells (CD8+ in the paper) will recognize the foreign peptide and kill the presenting cell. Obviously (?) viruses have been around a while and have ways to get around this to some degree. Or they wouldn't be here.
The way they worked the presentation in the paper was to put the influenza genes onto another virus, vaccinia. Vaccinia is the virus used to immunize for smallpox. It's pretty wimpy overall. They infect people with the recombinant vaccinia virus, the immune system responds against it, including the recombinant influenza proteins.
Yes, there's some concern about what this infection might do by itself. You wouldn't want to give this to someone with a severe immune deficiency for example. That's why you do trials.
In this specific case, the recombinant virus
shouldn't be in danger of being a 'super virus' because of the way vaccinia replicates vs the way influenza replicates. They're very very different viruses and the function of NP has little use in vaccinia replication.