Leonard Zon, Harvard Stem Cell Institute

As the longest and perhaps best-studied stem cell population, haematopoietic stem cells harbour conflicting data and ideas. In a review in Nature1, Leonard Zon of the Harvard Stem Cell Institute, in Cambridge, Massachusetts, explains how these cells can potentially be used to find a molecular definition of stemness.

Is there such a thing as stemness?

The original use of the term stemness was derived from a few articles about four years ago when several labs looked at genes expressed in stem cell populations and compared lists. Originally it was exciting, but when the lists were compared there were only a few, if any, genes that were truly shared. That made people question whether gene expression by itself could actually define stemness. Most people in the field now believe there is no such thing as stemness, at least as it relates to a unique set of genes expressed in stem cells to do a particular job.

One reason that you might not have a common signature is because of the diversity of genes that a vertebrate has within a certain family of factors. In a blood stem cell a Wnt factor involved in stem cell self-renewal might be Wnt3, but for a neural stem cell maybe it's Wnt1. The pathways may be common to stem cells, but the actual names of the genes may not be. That remains untested.

How can you have a short-lived stem cell if a stem cell is defined by its ability to self-renew?

There is no single stem cell in the blood; it's a diverse population of cells that has the capacity to self-renew. And the realization of that potential and the capacity of the cells is very much like that in school.

If you take a thousand kids who are graduating high school, they are all seniors, so say they are all stem cells. But their fate is different. So a short-term stem cell is one in which they go to college, maybe they renew themselves four times, and then they're done, they go get a job, they take on a cell fate. But then others, long-term stem cells, go on to graduate school.

So is it really like school, where a college senior might weigh graduate school and job applications at the same time?

It's very, very similar to an educational decision and how the environment actually regulates that process.

What are competing transcription factors competing for?

This is not incredibly well understood, if you take the case of the myeloid versus erythroid fate decision. Based on the work of Thomas Graf and others, if you have GATA1 and PU.1, they are actually binding to each other in cells. They form a complex; they pretty much neutralize [each other's] activity when they are complexed together, so it is only the free transcription factors that are available to bind DNA. So if you have an excess of GATA1, you become a red blood cell. If you have an excess of PU.1, you become a myeloid cell.

I was thinking that they would compete for spots on DNA.

That is one mechanism. I would say both are happening. If you consider those factors that you put in for pluripotency, they set up what someone has called a transcription factor battlefield, where the factors are actually complexing each other, inactivating each other. That ends up creating chaos that actually affects the cell fate.

Why is the notion of where blood arises in the embryo such an interesting question?

It's intrinsically interesting that stem cells are born at a certain time in development and then migrate to specific locations where they reside to wait for a developmental fate. It may be that way for other organs.

The latest thought on this is that all your stem cell precursors are made before day 12.5 in a developing mouse. After that period it's simply the derivatives of the stem cells that land in a niche and then are able to be amplified. So one question is why would you go to the trouble of making or maintaining stem cells in different locations for particular organs?

The second thing is that understanding blood development has helped us understand a lot about the genetics of haematopoiesis; the genes that are required to make blood cells in an embryo are often used in adults in the regulation of self-renewal.

It seems to generate a lot of controversy.

Pretty much every five years or so there is a challenge to the current hypothesis, and one would wonder, with the amazing tools we have in the blood system, why we can't just figure it out.

A third reason has to do with personalities. Pretty much every five years or so there is a challenge to the current hypothesis, and one would wonder, with the amazing tools we have in the blood system, why we can't just figure it out.

Stuart Orkin and I just wrote a review for Cell and as we were going through the field of haematopoiesis, we realized that many, many topics had conflicting data associated [with them], and it made even two experts in the field have difficulty in deciding what is the truth.

If you were writing a review called “Intrinsic and extrinsic control of haematopoietic stem cell self-renewal” ten years from now, what do you think it would say?

In ten years there will be a chart in the manuscript that would delineate the interactions that are necessary for self-renewal to occur. You could look at this as an electrical diagram as much as a computer program of cell fate and self-renewal. There's a growing literature in the haematopoietic system about how the transcription factor competition model allows you to think of cell fate. I'd like to see that put together schematically with the genes actually defined. A number of the signalling pathways are inducible, and I'd like to see how the inducible pathways interact with the pathways that are autonomously programmed, that are simply part of the cells' fate.

What will be some of the hardest things to put in this schematic?

It's going to be groups of factors that are truly making those decisions, and we need to be better at integrating the multiple factors and multiple signalling pathways. A number of labs pick a single factor, knock it out and show that it's relevant, but it may be that combinations of factors are what's required to lead to a biologic effect.

What else?

One thing to point out is the use of these pathways in tissue regeneration and repair compared to steady state. There are pathways that do regulate steady state (cKit and thrombopoietin), but other pathways like Wnt and Notch seem more relevant to the regeneration or tissue repair situation, so understanding similarities and differences between homeostasis and repair will give us a greater understanding of stem cell biology.


Cell fates and career decisions Credit: Leonard Zon