Well, that depends, whether you want to do temporal or permanent transformation.
If temporal, than you should transform cells, which are supposed to express the gene (you probably won’t transform muscles with gene for insulin;)
If permanent, than you have to transform germ cells and develop whole organism, which should be whole transformed.
This applies for multicellular organisms, but you will probably want to transform some one-cellular organism for mass-production of e.g. insulin.

in this post i am going to talk about multicellular organisms
you will want to create the transformation in the b cells of the islets of langerhaan
this can be done using bacterias or viruses

if the change is done in the germ line then the particular organism isn’t affected but its future generations are

you could also use liposomes but again that’s very hard

to make the change in germ line will be easier but still its considerably hard

if you make use of bacterial cells or vrius there are high chances of infection
and another thing you could insert a transformed cell in the area and it would divide over a period of time and give good results

another thing you could merge the transformed cell with hybridoma cells but there’s a risk of cancer. and that would need to be treated

another is idea i ain’t sure of its feasibility but is you remove all the B cells of islets of langerhaans and put a few transformed one

the absence of other will make the body initiate its division

I know the question’s already been answered but i just wanted to highlight this line and make clear just in case, that it’s NOT that all cells in our body produce insulin. As it is with all other hormones produced in the human body, each one is produced by a specific gland or organ (in this case, as jwalin pointed out, the pancreas).

Really? The question was about transformation of (human) cells. So, if I transform them with gene for insulin under control of some constitutive promoter, than only pancreas will produce it?

Sabbi is right, and you better reply to this thread so we know this isn’t some homework question. Once you are an adult, the standard gene therapies have this limitation of one cell only, or a few of cells. Assuming the gene insertion succeeds(more often they screw up) there is always the issue of not producing enough insulin just because you only have a few cell active. So does this mean gene therapy sucks, yeah probably, but of course there are now better methods to do these things than just inject someone with a virus and call it a day. Unfortunately, many of them again sucks, but they suck less relatively, and are only in the trial stages.

By the way Jwalin’s suggestion is something that has passed clinical trials so far.

Fine, i guess i’m wrong.