Have you ever asked yourself, “how is it that our immune system can fight off almost everything?”
It’s one of those things that is easily ignored.
It works. That’s all I care about.
If that’s not a good enough answer, then read on…
The answer lies somewhere between biology and statistics. And I want to start with an analogy.
Think of a website that makes you log in when you visit (Google’s gmail, for instance). You come up with a password when you join and then use it every time you log in. Some annoying websites make you cycle your passwords regularly for security purposes. (I’m not saying there’s anything wrong with that, but it can be taxing to those who don’t use a password storage program. – by the way, I use Dashlane and love it)
But every time you use a password, it’s off the list – you can’t use it again. So after about three changes, you start to sweat because you think your head is filling with all those old passwords and you can’t remember the latest one any more. As an analogy for the immune system, imagine a simple program that creates random passwords for you and ensures that they’re not repeats of any that you’ve used before.
Your immune system has no idea what you’ll come up against in the world. All it can do is make a vast repertoire of immune cells with the hope that it will be sufficient to react to anything. For simplicity, let’s just consider how cells make antibodies. To do this, your cells have a way of randomizing the protein sequence responsible for making these proteins.
The problem with this system, if it’s just a random grab-bag, is that sometimes those antibodies might bind to your proteins causing big problems. So, after the random process that generates antibodies, there is a second, non-random selection process that eliminates any that bind to you.
In my analogy, imagine that a random password is generated (the antibody), but then it checks to be sure it’s not the same as a previous password (no self – reactivity).
If you do any programming you can imagine outlining your code:
(let’s say passwords are 4-digit numbers from 0-9)
- generate a random number from 0001 – 9999
- cycle through old passwords
- check that the new password is not equal to the old password.
- If it matches, discard that password and go back to step I
- If it does not match, cycle to next old password
- Repeat until all old passwords have been checked
- Present new password to user
Now that I look at it this way, it is very much like evolution by natural selection. Random process à non-random selection.
To illustrate how this works with the actual proteins, it’s best to go to good old Janeway:
The top two panels show something more like the actual structure of the antibody. The bottom panel shows a simplified cartoon, highlighting the variable region and the constant region of antibodies. Think of the constant region as the backbone of the molecule – it comes in a few models, but doesn’t change.
The variable region is where the antibody binds its target. This is the region that gets scrambled up so the antibody will have a unique binding region.
The variable region is actually composed of several parts (V, D and J) that get pieced together, one of each sort. This accounts for some variability, but could only result in a handful of different types.
In addition to this mix-and-match, the joining of segments is also imperfect. Recall that DNA is ‘read’ in three-base codons. Because of this, adding one extra base in joining the elements will result in a frame-shift that creates even greater diversity. It also admits the real possibility that the protein made will be entirely unstable and useless. To account for this, each cell is positively selected for only ones that make stable receptors. It has two shots* at making this work. Once for each of the two chromosomes (one from your mom, and one from your dad) bearing this gene. If it succeeds, it goes on developing**; if it fails, it commits cellular suicide: apoptosis.
The result is a pre-immune repertoire of about 1012 antibodies available to protect you from any nasty ‘bugs’ out there.
* There is data supporting additional receptor editing.
** Heavy Chain is rearranged and interrogated first, then Light Chain.