The Loom

Before I head for Utah, let me direct your attention to two articles of mine in tomorrow’s New York Times. They don’t have a whole lot in common except they are examples of cool biology…

1. Virus traps. Here’s a case where ecology, evolution, and medicine all come together in an intriguing mix. You can think of any population of animals, plants, or other organisms as a leaky bucket under a running faucet. The population is boosted by sources of new individuals, and drained by sinks. Sources may include rapidly reproducing individuals, or immigrants from other populations. Sinks include the death of individuals in the population, or their failure to reproduce. Source-sink dynamics are important to ecologists, in part because the survival of a species may depend on sinks not draining its populations faster than its sources can restock them. Humans can create new sinks that tip this balance. If a forest is reduced to fragments, seeds that drift out into the surrounding farmland may be unable to grow, for example. Humans can even create so-called “ecological traps” that draw animals to their doom. Mayflies, for example, sometimes lay their eggs in parking lots because they look like rivers.

The same rules that govern mayflies also govern viruses. They have their sources and sinks, too. Their sources are the new viruses that spew out of infected cells, and their sink is the immune system’s attack. Some scientists are investigating whether they can add another sink: an engineered cell that lures viruses to it, but inside of which they cannot reproduce. (For more on virus traps, see this paper in the journal Ecology Letters.)

2. Marmoset chimeras. These cute little monkeys pose a serious philosophical quandary. Each individual marmoset is an amalgam of cells from two different marmosets. That’s because marmosets are generally conceived as fraternal twins and then trade stem cells like mad. Even eggs and sperm can belong to siblings. It may look as if a male marmoset is fathering baby marmosets, but he may actually just be their uncle. (Here’s the paper.)

I called up the Harvard evolutionary biologist David Haig to talk about this finding. In 1999 Haig wrote a speculative paper (pdf) about marmoset evolution. At the time scientists had only found chimerism in marmoset blood. But Haig mused about the possibility that marmosets might actually be more mixed. Now those speculations appear to be on the mark. Haig is interested in evolutionary conflicts of interest–when natural selection favors different strategies for two intimately connected organisms. Mothers and children, for example, may not get the same evolutionary benefit from a pregnancy–which may explain disorders of pregnancy such as preeclampsia. (I wrote more about this here.) The new marmoset research has Haig musing anew. In the article I only had room to quote him on a couple points, but he had plenty of other intriguing ideas.

One of those ideas had to do with the fact that marmosets are unusually prone to getting sick from viruses. This may have something to do with the fact that marmosets also produce relatively few receptors on the surface of their cells that the immune system uses to recognize them. Why would marmosets do something that was so harmful to themselves? It may have something to do with how marmosets become chimeras. The placentas of twin marmosets fuse, creating a network of vessels through which cells from one twin can travel to the other. Stem cells move back and forth thorugh this network. And perhaps immune cells do as well. Haig suggests that natural selection would favor twins that could attack their siblings in utero, sending out immune cells to wreak havoc. Reducing the receptors might be a way for a marmoset to avoid being recognized–and attacked–by the immune cells of its twin.

Update 3/27: Haig link fixed.


  1. #1 Corvus
    March 26, 2007

    That is amazingly cool.

  2. #2 Peter Ellis
    March 27, 2007

    Evolutionarily, it seems to me that the best way to regard the marmoset story is as parasitism. The embryos are conceived separately, but stem cells from one then invade the other in order to secure an extra route into the subsequent generation.

    That being so, you might predict a higher degree of chimerism in the testis than in other organs, since that’s the selective pressure driving the phenotype. And from your article it seems like that’s what’s observed.

    Not sure why they don’t also observe high levels of ovarian chimerism, though – possibly because if you get XY cells colonising an ovary, then a proportion of the next generation would be YY and thus nonviable?

  3. #3 jackd
    March 27, 2007

    Caught this on the BBC website this morning: Semi-Identical Twins Discovered. It’s a writeup of a Nature article on a set of twins in the US who are chimeric, but who have some XX cells and some XY cells. The conclusion is that one egg got fertilized by two sperm, which then split to form two embryos.

  4. #4 Marco Ferrari
    March 27, 2007

    Carl, there something wrong with this:
    “Reducing the receptors might be a way for a marmoset to avoid being recognized–and attacked–by the immune cells of its twin.” I was expecting an arm race, so maybe a higher number of immune cells to attack the twin, not a lower number of receptor to hide from it (him, her?). This way, you end up being attacked by viruses and parasites, apart from your twin.
    As an aside, what about armadillo (Dasypus I think)? They too reproduce, by quadruplet. Any study on the species?

  5. #5 Ross
    March 27, 2007

    >I was expecting an arm race.

    I’m not sure that an arms race is the best way to think about it, as the phrase in human terms tends to create an expectation that the party with the most arms will win (Vietnam and Iraq notwithstanding). A more neutral model is needed. A marmoset with more immune cell receptors than average may well be a healthier, more potent adult. However, to become an adult it must first survive being a foetus.

    Every foetus will have a slightly variable number of receptors, from more to average to less. If a foetus has a more than average number of receptors, he is more easily recognised as ‘other’ by his twin. So he is more at risk from his twin’s immune system. Such a marmoset is more likely to die in the womb as a result of an attack from its twin’s immune system. So a high number of receptors is of no evolutionary value, in fact it is a disadvantage, in the womb.

    It is possible there could be a see-saw effect whereby the disadvantage, in the womb, of having more receptors is negated or outweighed by an advantage, for adult marmosets, in breeding success. However, it appears that what actually happens for marmosets is that over time the impact in the womb has been more significant and that the average number of receptors in the population, compared with other mammals, has reduced over time.

    There is no long term direction to evolution, just a series of ad hoc solutions to environmental problems based on whatever relevant genetic variability is present at the time that improves the survival odds.

  6. #6 Marco Ferrari
    March 28, 2007

    “However, it appears that what actually happens for marmosets is that over time the impact in the womb has been more significant and that the average number of receptors in the population, compared with other mammals, has reduced over time.”
    I agree with you, Ross, it does make sense. But in most instances I know, the number of (put what you want here, from tine number to muscle weight to feather length) goes up, not down. And that’s because of the danger of going down, up to disappearance. I mean, a perfectly invisible foetus, without receptors, is a perfectly dead pup.
    I was just amazed by a solution not seldom seen in nature, the downgrading of weaponry, so to speak.
    Well, maybe marmoset are like bdelloid rotifers, apparently the only ones surviving without sex since tens of millions of years.


  7. #7 A S
    March 28, 2007

    Maybe I’m a bit confused. As I understand it, there are two genotypes, but each body has a mixture of cells from each genotype. If so, how can in-the-womb competition have much of a selective pressure? I mean, if both bodies A and B contain a mix of genotypes I and II, then how does twin A killing twin B favor one genotype over the other? [I'm assuming that the genitals don't take a big role in immune responses in utero]
    And, again, maybe I’m missing something (I’m not much of an immunologist), but it seems obvious to me that potentially having two different genotypes of immune cells floating around in one body, along with two different genotypes of non-immune-system cells will dramatically increase the risk of auto-immune disorders. And that there would be some selection pressure for reducing somehow the aggressiveness of the immune system.
    Of course, every species with an immune system needs to balance advantages of aggressive immune response (less disease), with the disadvantages (allergies, auto-immune disorders). But as I understand it, marmosets face an extra-high risk of the latter, so their optimal balance is for less aggressiveness.
    Am I missing something?

  8. #8 Ross
    March 29, 2007

    >if both bodies A and B contain a mix of genotypes I and II, then how does twin A killing twin B favor one genotype over the other?

    Although bodies A and B contain a mix of genotypes I and II, it isnt an equal mix. So it is entirely possible for one genotype to prevail over the other and for its dominance to build up over a series of pregnancies. It will usually be a majority/minority result rather than an all or nothing result . In his paper, Haig refers to evidence for ongoing diversifying selection, so we would expect to see the dominant genotype showing increased numbers of receptors over time. Haig then refers to a paper (Cadavid et al) that found periodic turnovers cause the descendants of one sequence to replace all the others.

    I would interpret this as meaning a point is reached where the dominant geneotype has so many receptors that its offspring become vulnerable in the womb to the antibodies from the other genotype. So there is a sudden changeover from a majority of genotype I to a majority of genotype II. I assume that over time genotype III will evolve from II or genotype I will lose potency and the cycle will repeat. So the marmoset evolution will alternate between a few generations of competition pressure between foetuses in the womb, which leads to fewer receptors, followed by a few generations of competition pressures between adults, which leads to more receptors. The long term trend has been towards fewer receptors.

    Rightly or wrongly, I have mental pictures here of Galapagos finch average beak sizes getting larger or smaller from one year to the next depending on the food sources available, or wolf/moose numbers fluctuating on Isle Royal in Lake Superior. I think at a very general level we are seeing similar genetic pressures across all these examples.

  9. #9 A S
    March 29, 2007

    OK, first, am I still missing something about immunology that explains why chimeras wouldn’t be more at risk for auto-immune problems?

    And for in-womb competition to be a factor, the advantage of using your immune system to kill your wombmate would have to be greater than the disadvantage of immunologically attacking large parts of your own body (and that’s before discounting the possibility that either you or your wombmate have chimeric gonads).

    In fact, thinking about it, it seems at least plausible to me that the chimerism is a result of decreased immune system sensitivity rather than vice-versa (although perhaps my ignorance of immunology, embryology and other -ologies will show here). Assume an ancestral marmoset similar to other mammals (no chimerism and typical immune systems), except that fraternal twins are typical, and for some reason the anatomy of marmosets leads to mingling of placentas (and therefore mingling of antibodies, etc. among the fetuses). With typical immune systems, that would often lead to fetal cells being attacked by antibodies from its wombmate. So there would be pretty strong selective pressure for cells to avoid triggering immune systems from other genotypes. After a while, the immune-recognition of cells could have been low enough that stem cells that accidentally migrated from embryo A to embryo B wouldn’t be recognized as foreign by embryo B. Bingo! chimerism.

    Any real experts in the relevant –ologies, please help me understand mistakes in this theory.

  10. #10 Kilian
    March 29, 2007

    What I’m wondering, having read the article about the human chimeras – how do the marmosets avoid having many hermaphrodite individuals? Sperm of a male marmoset may be that of his brother’s, but what if he has a sister? Do sterile marmosets exist because of this? And what about behaviour – can different behaviour be detected amongst marmosets with both xx and xy genes compared to normal males and females?

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!