There’s another really cool part of the story: antigen “encryption” by the immune system.
In addition to the antibodies produced by B-cells, part of the adaptive immune response is mediated by T-cells that recognize antigens through receptors on their surface. These T-cell receptors have the same range of random diversity in antigen specificity (generated by V(D)J recombination) as antibodies. However, T-cell receptors do not recognize “raw” antigen. They recognize carefully processed antigenic peptides that are randomly chopped up and “presented” to them on the surface of Antigen-presenting cells. These peptides (pieces of antigen, whether self proteins or chopped up parts of pathogens) are held in place on the surface of Major Histocompatibility Complex (MHC) cell-surface proteins, and the epitope (shape) that the T-cell receptor recognizes is the complexed shape of the antigenic protein plus the MHC protein that’s gripping it. In effect, the antigen is encrypted by the MHC molecule when presented to the T-cell receptor.
Now, the MHC region of the genome is one of the most diverse (polymorphic) in the genome - i.e., there are a wide variety of MHC “haplotypes” present in the population. Here, we are talking about hard-coded DNA variation, different people in the population carry different variants at this locus in the genome, and pass on their particular variants to their children.
What this means is that two individuals with different MHC haplotypes will “encrypt” the exact same pathogenic antigen in different ways, to produce a different epitope (shape) for the T-cell receptors to look at. All of this makes life much harder for a pathogen. Even if a pathogen can learn to “display” only epitopes that look similar to host proteins on its surface, the pathogen’s proteins will be chopped up endless different ways, and then encrypted in different ways by different people in the population according to each person’s particular MHC haplotype.
MHC “encrypted” presentation of antigens is the reason for the need to tissue-type people, because the host proteins are also encrypted in the same way during the process early in life when the immune system learns tolerance for self. Self-antigens that are encrypted by a different MHC haplotype will be treated as foreign. Thus, a “tissue match” is somebody with a similar MHC haplotype.
The other interesting thing about the MHC region is that there is obviously an evolutionary advantage to having a different MHC haplotype from other people in your local population. If pathogens have not yet encountered your particular encryption key, they are unlikely to have evolved to subvert it. Thus, the MHC region is subject to strong balancing selection, the unusual form of natural selection that favors diversity.