Blood type matching

I've been swotting up on five different techniques used in immunology to detect various proteins in blood, hair, food and other samples. This is because I have to do a test on them next week, and mostly it's just been a question of looking through the lectures and making notes to help me remember. Towards the end of the day I had a look at what a couple of text books have to say, and this is where I came upon a little tiny snippet of interesting information, about blood typing.

This is about how they work out the different blood groups: A, B, AB and O. The A and B labels relate to two proteins on the surface of red blood cells. People with blood group A have A proteins on their red blood cells, people with blood group B have B proteins on their red blood cells, while people with both are AB, and if neither is present then they are group O. So far, so good.

The basis of working out which you are is down to the immune system, which is finely tuned to determine 'self' from 'non-self' based on various immune cells in your body coming into contact with other cells, proteins, viruses or whatever. This is done intensively at the start of life, so that anything that reacts with 'self' is targeted for destruction - one of our other lectures covered this but I haven't reviewed it since, so I'm a bit vague on how that's done. What you end up with is a trillion zillion antibodies that recognise anything that isn't you, based on what your immune cells have interacted with up to that point in life.

So, people with blood group A have antibodies that recognise protein B, because it isn't A, so indicates 'non-self' and needs to be zapped. Similarly, people with blood group B have anti-A antibodies, people with group O have anti-A and anti-B, and AB don't have either anti-A or anti-B. Are you keeping up? Nearly there.

What I never thought about before was why group O people always have antibodies to A and B when they might never have come into contact with A or B proteins? And similarly, why group A always have anti-B and B's always have anti-A? It turns out (according to the textbook) that common gut bacteria have proteins that are similar or identical to A and B, so they stimulate the formation of the anti-A and anti-B antibodies. What this doesn't explain is why the bacteria don't cause trouble in AB people who don't have antibodies to A or B - presumably they aren't harmful bacteria.

I'm pretty sure that the immune system can generate random antibodies for things it hasn't come into contact with yet, so maybe this bacterial explanation isn't the full story. But it provoked a glimmer of interest in a pretty dull day. The language of immunology is impenetrable, and there are millions of factors with indistinguishable names, acronyms and abbreviations.