What I taught today

I was bad. I didn’t post my summary last week, so this is actually what I taught a week ago and what I taught today.

Previously, I’d given an overview of the foundations of modern developmental biology in embryology and anatomy. I gave them more history last week, only not so ancient: what led to the modern focus on patterns of gene expression? In the early 20th century, it was all Entwicklungsmechanick, the experimental manipulation of embryos and analysis of morphology in order to infer mechanisms of transformation (that’s a bit of an oversimplification: we were also interested in physiology, and some vital dyes, for instance, also let us see patterns of metabolic activity). But all that was indirect. We were looking at the output of a black box, but we couldn’t see what was actually going on inside the black box, where all the action was. And everyone knew it. Here’s what TH Morgan said in 1927:

“One of the most important questions for embryology relating to the activity of the genes cannot be answered at present. Whether all genes are active all the time, or whether some of them are more active at certain stages of development than others, are questions of profound interest.”

Think about that. In the early days of developmental biology, we didn’t even know whether there was differential gene activity or not; it was considered a reasonable possibility that all the genes were just doing their work, whatever it was, all the time in every cell, and that differences between cells emerged farther downstream, in biochemical interactions. But they knew this was an important question. They knew that we had to look at the activity of individual genes…they just didn’t have the tools yet. So it was back to hacking up embryos and trying to infer causes from aberrations.

The change emerged gradually, but there were a couple of watershed moments where everyone looked up and noticed that hey, we do have ways of looking at genes directly. One was the work of Ed Lewis, a most excellent geneticist who used the tools of genetics to look directly at mutations that caused changes in fly morphology, in the 1960s. This was amazing stuff — the papers he wrote were beautiful and complex and very, very genetical — but it was written in a language that most developmental biologists of the day were unprepared to read. They were genetics papers. But I think they laid a foundation: if you want to do development, you’d better learn about genetics.

The second big event was the saturation mutagenesis screen of Christiane Nusslein-Volhard and Eric Wieschaus, about 20 years later. This work was also built on an understanding of genetics, but also used the tools of molecular biology. It was another lesson: if you want to do development, you’d better learn about molecular biology. I defined this field for the students:

Modern developmental biology is the study of differential patterns of gene expression and their effects on cells, tissues, organs, and organisms.

I also talked about the Evo-Devo ‘Revolution’, but I put a different twist on it. This is usually expressed by developmental biologists as the new change in thinking, the one where we’re going to teach everyone else how important development is to the field of biology as a whole, but I see it as something completely different. Evo-Devo is the discipline in which developmental biologists finally woke up and started paying attention to all the work other biological disciplines have done — we aren’t transforming them, they’re revolutionizing us by giving us the great good gifts of genetics and molecular biology to finally allow us to directly look at genes and gene interactions.

(I didn’t talk about this today, but will later in the course: another discipline, Eco-Devo, is the field where we also start respecting the importance of ecology and population genetics to development. Developmental biologists aren’t the masters, but the students — this field is so great because we’re benefitting from the synergistic incorporation of so many other fields. Just some of us are a little too cocky about it, and have a slightly skewed perspective.)

Slides for this talk (pdf)

So now I’m up to today’s talk, which is less history and more laying down common definitions so we can talk to one another like developmental biologists. We have our own language, with words with very specific meanings, and sometimes they are the same words we use in colloquial speech but with very different implications. I blamed this in part on Spemann, who was strongly influenced by the emerging literature in experimental psychology, and a lot of developmental terms have their roots in psychology…but have absolutely no cognitive implications! A conversation in developmental biology can sometimes sound very weird to outsiders.

So basically today I marched through a series of concepts and laid them out for the students.

Germ and somatic cells. A fundamental distinction between cells in multicellular organisms is that some are fated (and I have to define that word, too!) to become reproductive cells, or germ cells, cells that can contribute to future generations, and others will form the somatic tissues, or everything else. From a genetic perspective, everything is about the germ line, the population of cells that can pass on their genetic complement to another generation, and the somatic cells are simply support cells that help the germ line accomplish their goal. We’re all big lumbering organic capsules that work to protect and propagate a tiny clump of cells in our testes or ovaries — but from a developmental perspective, a lot of really cool and interesting stuff goes on with the somatic tissues. That’s where cells make brains or pancreases, while germ cells are often segregated early and don’t make decisions quite as elaborate.

Germ layers. 19th century biologists seemed to love the word “germ”, and it shows up all over the place. It just means a small mass from which more parts or organisms can develop; think about wheat germ, for instance, or germination. One of the simplifying concepts of development is to recognize the early segregation of somatic cells into distinct tissue layers, the endoderm, ectoderm, and mesoderm. One way to think of this is that the goal of organismal development is to produce an elaborate 3-dimensional organic structure from a linear, one-dimensional sequence of molecular information, and unsurprisingly, one early step is to make a set of 2-dimensional sheets. These sheets then interact with one another to generate a more complex 3-dimensional arrangement of cells.

Mosaicism and regulation. I repeated myself a bit and reminded them of the experiments of Roux and Driesch, and also introduced them to the beautiful work of Conklin on tunicates. One simple way to imagine development from one cell to multiple cells and cell types is to compare it to a paint-by-numbers project: the mother prespecifies the fate of each portion of the embryo by inserting spatially localized instructions in the egg. That means that each region of the egg is locked into a future role: one part may have molecules that compel it to form muscle, another might by packed with molecules for brain development. This is mosaicism: the egg is a mosaic of cytoplasmic determinants. An alternative pattern is to have an egg that is a blank slate, with only generic potential to divide and become any tissue of the body. In this case, different tissues arise by cells communicating with one another, and regulating to make the necessary parts. Two cells may talk to one another, for instance, and negotiate which one will make muscle, and which will make brain. Mosaicism and regulation represent two ends of a spectrum, and all organisms exhibit both to some degree. Conklin’s tunicates are very mosaic, while human embryos are highly regulative.

Organizers/Inducers. The work of Spemann and Mangold engaged embryologists in an often repetitive search for decades. They found that certain tissues of the early embryo had organizing activity: when transplanted anywhere in the embryo, they would induce or recruit neighboring tissues to form complex structures. Mangold’s experiment, for instance, involved transplanting a site in the amphibian embryo where gastrulation began called the dorsal lip, and finding that wherever it was, it would initiate gastrulation and lead to the formation of a complete body axis. As it turned out, lots of transplanted substances could induce whole body axes; this was a clue that the organizer wasn’t delivering detailed instructions, but was a source of very simple information that would trigger a cascade of events that led to complex consequences.

Fate/Specification/Determination. These are misleading terms! We use them all the time. “Fate” just refers to the likely final differentiated state of a cell or tissue in a particular position. For instance, a cell located at the animal pole of an embryo might be very likely to develop into the nose of the animal because of its position, but having that fate may not imply that it is in any way molecularly predetermined to end up there. A child born to Mitt Romney, for instance, is very likely to end up rich and secure; it does not mean that it was born with genes for richness. We often define fate operationally: we label cells in particular locations in the embryo, and ask where they end up in the adult.

A cell is determined (that psychology source is showing up again, isn’t it?) if it is actually locked in to a particular fate. Again, this is operationally evaluated: if you transplant the cell from one place to another, if it develops according to its destination location, it is not determined; if it develops according to its source location, it is determined. So if I pluck out an early embryonic cell from a location fated to become the eye, and put it in a place fated to become the tail, and later observe an adult with an eye in its tail, I would say that the cell was determined to become eye tissue at the time of the transplantation.

Specification is an intermediate step. Say I transplant a cell fated to become eye to the tail, but in the adult it is transformed into normal tail tissue. It is not determined. But let’s say instead that I transplant that cell to neutral ground — an empty petri dish — where no neighboring cells can influence its development, and it goes on to make eye tissue. It is not determined, but I can say that it is specified to make eye.

These terms and experiments were formulated at a time when the role of genes was unknown (see that earlier TH Morgan quote!), but they make even more sense now. What’s going on in development is a stepwise, progressive pattern of fixation of gene activity, turning off some genes and turning on others, that leads to differentiation of a particular cell type.

These are important concepts for understanding modern stem cell research. The goal is to have a set of totipotent cells (cells that have the potential to become any cell type in the adult) and understand the patterns of cell signaling/communication that induce a specific sequence of gene activity that transform the unspecified cell into a specific cell type.

Slides for this talk (pdf)