Our textbook, and how to learn the most from it:

In order to understand our textbook, and make best use of its virtues, imagine a history of the American Civil War, that was written by a Confederate General while that war was still going on.

Such a book would have deep-seated assumptions, that it would regard as profound insights. In some cases these really would be deep insights. Other important facts would systematically never be mentioned. For those who can "read between the lines" such a book could be a better source of wisdom than alternative histories that tried to be unbiased.

Anatomy consists of certain spatial arrangements of differentiated cells: bone cells here, muscle cells here, bone cells there, brain cells here, skin cells there, etc. for about 250 cell types in mammals, but only for 5 or 10 cell types in sponges, hydra and other simple invertebrates. The slime mold Dictyostelium has 4 cell types: amoebae, spore cells, stalk cells, and a cell type that forms the base of the stalks.

The central question of embryology is to explain the mechanism(s) that allocate undifferentiated cells to their different fates (organs and differentiated cell types). A major part of this question is how the mechanism adjusts in proportion to the volume and number of cells to be allocated.

Specifically, if an embryo is half as big, then all the organs are (to begin) made half as big; when an embryo is twice as big, then the parts are made twice as big. This "regulation" doesn't always succeed; and development is much more "regulative" in the embryos of some taxonomic groups than others.

Mammal embryos are the most "regulative", in this sense, of any kind of animal that I know of. Sea urchins and other echinoderms are about in second place. Frog and salamander embryos are also regulative, but less so than echinoderms. Nematode worms, sea squirts and Drosophila (fruit flies) are at the other extreme. If you separate their first two cells, those cells will develop into halves of a body, with lots of organs missing. The word "mosaic" is used to refer to non-regulative embryos.

It is not agreed what underlying, fundamental differences in embryological control mechanisms cause development to be regulative in some taxonomic groups and mosaic in others. Many embryologists do not regard this as a fundamental difference, but just a matter of how early in development cells become irreversibly committed to their fate (i.e. which cell type they are going to differentiate into). Early commitment-> mosaic development; Late commitment -> regulative development.

Another belief (expectation, prediction, theory) is that the difference is fundamental, more specifically that when cell fate is controlled by special cytoplasmic components being put into certain cells early during cleavage, and controlling what those cells differentiate into, then that causes development to be mosaic. In contrast, development would be expected to be regulative to the extent that differentiation is controlled by chemical signals traveling from one cell to nearby cells (which is called "Induction", in the embryological sense of the word induction, somewhat on analogy to how this word is used in relation to magnets inducing electric currents).

However, as UNC students, you should learn that Prof. Goldstein is famous for discovering an example of embryonic induction in Nematode worm development (and proving this elegantly and conclusively). This surprised a lot of scientists, who had assumed nematodes would use (only?) cytoplasmic segregation, and (never?) use induction.

Personally, I used to joke that part of the attraction of nematodes as a model organism is that they aren't capable of regulation – because scientists couldn't discover the mechanism of regulation, therefore they switched their studies toward a kind of animal that isn't capable of doing these unexplainable things.

This is not entirely a joke. Embryonic regulation was discovered by Hans Driesch (1867-1941, but you don't need to memorize these dates). But do learn 1895, which is the year Driesch discovered regulation in star fish embryos, by shaking water with embryos in it. Notoriously, Driesch concluded that no possible mechanism can explain regulation, and therefore eggs must contain supernatural, mind-like entities that (who) notice damage and make conscious changes to compensate and fix things. He used the word "entelechy" to refer to these imagined mind-like entities, and switched from embryology to parapsychology. Better he should have switched to nematodes! They are unable to do what he couldn't explain. At least Driesch didn't just ignore facts that contradict his expectations, like our textbook does.

More attention should be given to an earlier experimental result (Roux, 1885) that if you use a red-hot needle to poke cells of early amphibian embryos (which actually didn't quite kill them, as Roux describes in the original [German] paper), then regulation fails to occur. An English translation of the original German paper hass been published in an anthology. In other words, if you remove one of the first two cells, then the remaining cell regulates; but if you "kill" (severely damage and slow down development) one of the first 2 cells, and leave it in place, then regulation doesn't occur. Somebody ought to repeat Roux' experiment really killing cells. A UNC undergrad and I tried to repeat this experiment 20 years ago using Xenopus 2 and 4 cell stage embryos, but it's a lot more difficult than we expected. For example, the frogs only laid eggs in the middle of the night, mostly on weekends. Also, the whole embryo always died at about the 16 or 32 cell stage. I think sea urchin embryos ought to be used instead. We got very good at poking cells with tiny red-hot needles.

Two unexpected facts that textbooks rarely mention, because they don't fit anybody's theory:

If you cut off the end of a pluteus' arm, its cells will form a hollow ball. Not only does this hollow ball of cells look like a blastula, it gastrulates, and forms a whole new pluteus (a rather small, "scale model" pluteus, with all organs formed with proportionally smaller size).

Many species of wasp lay their eggs inside the bodies of caterpillars. Surprise #1: These egg cells have holoblastic cleavage, very much in contrast to the syncytial, 6000 nuclei in one big cell that Drosophila embryos develop into. Unorganized masses of cells grow in the body cavity of the caterpillar,

and (Surprise #2) clusters of undifferentiated wasp cells aggregate and (somehow!) differentiate into wasp bodies, sometimes hundreds of them, from one original fertilized egg cell. This is called polyembryony. Nobody knows how to explain it. It is extremely regulative development, in contrast to the mosaic development of flies.

Cell Differentiation: What causes particular subsets of genes to become locked "on"? A possible explanation would be for all the genes that are turned on in pancreas cells to have the same base sequence just upstream of them, analogous to how operons work in bacteria. For each of the 250 differentiated cell types, there could be a special "cis element", promoter region kind of control site for RNA polymerase to bind to. Maybe that's really how differentiation works. Evidence is over-due to be discovered. I will be perfectly happy if cell differentiation works that way. I would also be happy if all ectodermally-derived differentiated cell types had some special cis DNA sequence, and a different special sequence for all genes expressed in mesodermally-derived cell types. Scientists have been looking for patterns like this for a long time, and just because they can't find it doesn't mean that differentiation works some fundamentally different way. But it does mean I will withhold my applause until they discover the mechanism.

Not enough attention (Any!) is given to the mutual exclusiveness (for lack of a better name) of cell differentiation. Becoming a liver cell means being unable to be a pituitary cell, etc. etc. for almost all cell types. Methods exist by which you can fuse 2 cells in tissue culture. You can fuse a heart cell with a pancreas cell. You can fuse a mouse cell with a human cell, or with a chicken cell. Here is an interesting part: if you fuse a mouse liver cell with a chicken liver cell, then the fused cell will continue to make pancreas-specific proteins (both mouse proteins and chicken proteins). The same is true for other differentiated cell types. But if you fuse a liver cell with a skin cell, the fused cell will stop making pancreas-specific proteins, and stop making skin-specific proteins. (Actually, this hasn't been tried for all combinations of all 250 different cell types.) Here are two interesting exceptions to the general rule:

If you fuse a chicken red blood cell with a mouse pancreas cell, the resulting cells will start making chicken pancreas proteins, and continue to make mouse pancreas proteins. The reason is believed to be that red blood cell nuclei have "turned off" expression of all genes, making them susceptible to being converted to transcription of genes for (probably any) different cell types. Mammal red blood cells through out their nucleus during differentiation. But this is special to mammals. Shrunken, inactive nuclei are still present in red blood cells of birds, reptiles, amphibians & fish.

The second special exception is what happens if you fuse a mouse pancreas cell with a chicken skeletal muscle cell (which has dozens or hundreds of nuclei, in the same cell). The fused cell will start making mouse muscle proteins, in addition to chicken muscle proteins. You can do this experiment in reverse. The dozens of muscle cells somehow overwhelm the single pancreas cell, apparently.

There are other interesting special cases, including purification of a certain transcription factor that can convert just about any other cell type (including nerve cells!) so that the make muscle-specific proteins.

In principle, it would seem like a good way for embryos to create anatomical patterns if there were spatial signals (like those of a "Global Positioning" device) that told each embryonic cell which cell type to differentiate into. Another analogy would be to people in football stadiums holding up big pieces of colored paper, so as to spell "GO, HEELS". North Koreans excel in making pictures this way. If North Korea has textbooks in embryology, I expect they should be strong supporters of "Positional Information".

Furthermore, it would seem as if a good way to tell all the cells where they are located would be by means of diffusion gradients. There might be three diffusion gradients, each perpendicular to the other two. In addition to telling each cell where it is, such gradients might be able to explain regulation. For example, if you cut an embryo in two, the diffusion gradient would become twice as steep. Therefore, a half-size liver would develop, etc, etc.

The trouble is that actual embryonic development isn't anything like that.


back to index page