The selective pressure of childbed fever

A new paper in PLoS Biology by Xiaoxia Wang and colleagues finds that a gene called CASPASE12 (OMIM) has been under recent selection in humans. The twist is that the allele with a selective advantage is actually a null allele -- it doesn't produce any functional protein.

The substitution of a nonfunctional allele for a functional one produces a "pseudogene" -- a sequence with the form of a protein-coding gene, but that doesn't actually produce a protein. So the authors discuss this as a case of "pseudogenization", and trace the reason to one of the functional consequences of the loss of protein: increased survivorship from sepsis.

Our population genetic study provided strong evidence that the nearly complete fixation of a null allele at human CASP12 has been driven by positive selection, possibly because it confers resistance to severe sepsis. CASP12 is a functional gene in all mammals surveyed except humans [34], suggesting that it is indispensable in a typical mammal. The functional human CASP12 acts as a dominant-negative regulator of essential cellular responses including the necrosis factor-kappaB and interleukin-1 pathways; it attenuates the inflammatory and innate immune response to endotoxins [34]. Because an appropriate level of immune response that is neither excessive nor insufficient is important to an organism, one can imagine that the immune suppression function of CASP12 becomes harmful when the immune system cannot fully respond to a challenge. It is likely that during human evolution alterations in our genetic and/or environmental background resulted in a malfunction of the immune response to endotoxins, which rendered the previously necessary function of CASP12 deleterious in humans and the null allele advantageous over the functional one. Identification of such genetic and/or environmental alterations will be valuable for understating human-specific immune functions.

Sepsis is infection of the blood. It is most often caused by one of two bacteria: Staphylococcus and Streptococcus. These bacteria are near-ubiquitous, but each has pathogenic forms that are really nasty if they get into your blood. Drug-resistant forms are among the leading causes of death in hospitals today -- most of what you've heard about the problem of antibiotic resistance is directed toward finding solutions to hospital-related sepsis.

Of course, sepsis is never good. But drug-resistant bacteria in hospitals haven't been around very long. So we have to consider what used to cause sepsis, before hospitals did. Sepsis is a complication of many pathogens and wounds -- infections run wrong progress first to sepsis, then to death.

One form of sepsis strikes me as especially bad, and possibly especially common in history and prehistory: puerperal fever.

Puerperal fever used to be a common complication of birth. I like this intro from iVillage:

It turned maternity wards into morgues, but today few pregnant women have even heard of puerperal fever, an aggressive infection of the reproductive system following childbirth. It probably changed the course of English history by killing Jane Seymour, third wife of Henry VIII. It destroyed Isabella Beeton and ruined her husband's life and business. It finished off proto-feminist Mary Wollstonecraft, leaving Mary Shelley motherless.
In the 1700s, when it was known as childbed fever, puerperal fever could claim the lives of as many as 20 per cent of new mothers and would sweep through communities in terrifying epidemics. The infection - most commonly the bacteria staphylococcus and streptococcus - was often carried on the dirty hands and medical instruments of doctors and midwives.

For a long time, women (especially urban women) who sought out doctors to attend births had a way of ending up dead. One doctor might visit dozens of women, killing them all. During this time, women who lived in the country, were delivered by midwives, or just couldn't afford a doctor paradoxically did much better.

This is one of those things that makes you wonder why it took so long to come up with the germ theory.

Although it had its greatest incidence in the crowded cities of Europe 200 years ago, childbed fever was an omnipresent risk of birth throughout human history. Strep and staph are everywhere, and from an immunological perspective, women are uniquely vulnerable to sepsis after birth. Put the two together, and you have one of the major causes of mortality happening at a time that kills both mother and child.

So sepsis -- particularly this kind of sepsis -- was common and severe in terms of mortality. An allele that protects against it ought to be strongly selected. And although Wang et al. estimate the selection coefficient from sepsis survivorship data, seelction at the time of birth in reproductive age women will have more of a fitness effect. According to the study, there may be at least two others besides CASPASE12:

Pseudogenizations of the two paralogous MBL1 genes in humans and the finding that deleting Mbl1 increases survival in a mouse model of sepsis suggests that the losses of the two human MBL1 genes may have also been driven by positive selection. The common connection to sepsis among CASP12 and the two MBL1 genes reinforces the conjecture that the way humans respond to sepsis and/or the threat of sepsis to humans might have been significantly different from those in other species.

One of the aspects of CASP12 variation noted in the study is the near absence (total in their sample) of the functional allele among non-Africans, and its presence among 12 percent of Africans in their sample. I think it's too early to make much of this, considering the limited samples -- is it in Australia, for instance? They interpret it as meaning that an out-of-Africa bottleneck made it more likely to be fixed outside of Africa. I would read the opposite conclusion -- perhaps the allele originated outside of Africa and hasn't swept to fixation within Africa yet. In any event, it points out two things -- dating these events is tricky because you need to know about selection intensity if you don't know about recombination frequency, and some of these recent selection events are near fixation. The not-yet-fixed is one reason for differences in selected variants between populations, the too-new-to-spread is another.


Wang X, Grus WE, Zhang J. 2006. Gene losses during human origins. PLoS Biol 4:e52. DOI link