There is one question I have been asking me for a long time concerning Gene Therapy
Suppose you find a technique that allows you to introduce a gene product/protein in a patient that lacks it.
(e.g. Plectin or Dystrophin in Muscular Dystrophies)

Would this product not be "foreign" and trigger an immune response? Then the patient would end up with an autoimmune disease....

That's an insightful point, and immune response could be a problem in some senarios.

However if we look at some of the top targets for Human Gene Therapy:

*Cystic fibrosis caused by a defective chloride channel protein
*Muscular dystrophy caused by a defective muscle protein -dystrophin
*Sickle-cell disease caused by defective beta globin

They are caused not by the absense of the protien, but because it contains a mutation - often in only one amino acid, as the body tollerates the defective protein it is likley the newly introduced protein will also be tollerated.

In some applications of "Gene therapy" type applications inducing an immune response is the aim - for example in cancer therapy and DNA vaccines.

As Richard Taylor pointed out, dysfunctional proteins (owing to mutations) - usually in single gene diseases are currently the prime candidates for gene therapy.
Research over the last three years or so has shown that proteasomes (not just immuno-peoteasomes) continuously process damaged or misfolded proteins, and many of the peptides generated therof (DRiPs) are presented on the MHC, with no adverse immune reaction - in fact, as far as the immune system is concerned, its a learning experience. (I think there was a very good Nat immunol rev called DRiP-feeding the immune system).
In many kinds of cancers, non-conventional therapy (not necessarily gene therapy) is aimed at boosting the immune response against tumors that often actively try to suppress it. This can be done modifying presentation of antigens by APCs or even boosting positive costimulation, and downmodulating negative ones.