A recent paper on the origin of sickle-cell was being discussed, and an interesting question was posed [ by Razib Khan]: why haven’t modifier genes emerged that reduce the bad effects in homozygotes?
Let me say something about those bad effects. The most important one is that someone with sickle-cell disease usually has a ruined spleen before the end of childhood – infarcts. This enormously increases the risk from encapsulated organisms like pneumococcus: in a place or time without effective medical care, this kills you. There are all kinds of other problems that eventually show up if effective medical care gets you over that first hump, but in the past, and in many places today, kids with sickle-cell diseases usually didn’t make it past two.
What about a mutation that led to a milder form, one with the same positive effects in heterozygotes but that that killed homozygotes at age five, instead of two? It would not be favored by selection: five year olds don’t produce any offspring, and they use up more parental resources. Killing homozygotes rapidly (as opposed to slowly) increases the fitness of the sickle-cell allele – it lets the parents start over and produce another kid, who might be a carrier.
There’s a malaria defense in Southeast Asia, Melanesian ovalocyosis, that goes all the way – homozygotes die before birth.