You’ve heard of stem cell research and its promise of a medical revolution given the regenerative abilities of stem cells. But as it turns out, identifying what a stem cell is experimentally is not at all straightforward. Stem cells have two main abilities: cell renewal (division and reproduction) and cell differentiation (development into more specialized cells). The main problem is, there is no way to experimentally test whether one particular cell can both self-renew and differentiate to make more developed kinds of cells. Much like Heisenberg’s uncertainty principle, according to which we can’t measure a particle’s velocity and position at the same time, we can’t measure both properties that constitute a stem cell. Claims that any single cell is a stem cell are therefore inevitably uncertain, argues Melinda Bonnie Fagan.
The term “stem cell,” like so many others in biology, was introduced by Ernst Haeckel. In Haeckel’s speculative vision, an individual organism’s development (ontogeny) mirrors Darwinian evolution of diverse species (phylogeny), both processes emanating from a single “Stammzelle.” So the term originally referred to the root of the Tree of Life, for all species and for each living organism. The meaning of “stem cell” today is somewhat different. For one thing, the evolutionary aspect is has been jettisoned. Development is no longer thought to mirror evolution, and the two are now studied (for the most part) separately. Also, since Haeckel’s day, laboratory research on organisms, cells, and genetic material has generated enormous amounts of knowledge about development, especially in “model organisms” like fruit flies and inbred mice.
Yet for all this hard-won accumulated knowledge about the internal workings of cells and other organismal parts, stem cells remain hard to pin down. While today’s concept is less speculative than Haeckel’s, entities corresponding to that concept remain hard to identify, manipulate, or tame for routine medical use. The reason for this is both conceptual and experimental. Briefly: stem cells are defined in a way that makes them tricky to experimentally identify. The root of the difficulty – the stem, to be precious – is that what counts as a stem cell changes with biological and experimental context. Stem cells are context-dependent.
The cell types that a stem cell produces are more specialized, further along in the developmental process, than their stem ancestor.
Stem Cells in context
Stem cells today are defined by having two abilities, one reproductive and one developmental. The reproductive ability is “self-renewal”: a stem cell can divide to produce at least one other stem cell. The developmental ability, on the other hand, is “differentiation” – meaning, roughly, biological development at the cellular level. A stem cell can develop into one or more other kinds of cell. But not just any kind of cell. The cell types that a stem cell produces are more specialized, further along in the developmental process, than their stem ancestor. For example, a neuron is a specialized cell type with distinctive shape, internal organization, and patterns of gene/protein expression. Skin cells (fibroblasts) are another cell type. All multicellular organisms start from a single cell. In early development, a chain reaction of cell division produces many cells from the first. These early cells then begin to change, following diverse pathways leading to mature specialized cell types. An organism’s various integrated parts, its organs and tissues, are made up of those different cell types – ranging from a few to hundreds, depending on the species.
Differentiation is the “cell’s-eye view” of an organism’s development. In this way, the idea of a lineage tree persists in today’s stem cell concept. But the tree implied by today’s stem cell concept is not made up of branching patterns of species’ evolution. Instead, it’s a lineage of cells within or derived from one multicellular organism. An individual organism’s development begins with one cell and ends with the organism’s death. (Some developmental theorists define the process more narrowly, from one cell to reproductive maturity.) For multicellular organisms, the process of development involves increasingly specialized and diverse cells reproducing (by binary cell division; i.e., mitosis) and differentiating to form all the different kinds of cell that make up the body of a healthy, mature organism. So stem cells are context-dependent in a very concrete biological sense. Any stem cell is a stem cell relative to some individual multicellular organism.
A stem cell, by definition, participates in processes of cell reproduction, cell development, organismal development, and organismal regeneration. A lot of biology packed into a deceptively simple term!
Given this tie to multicellular organisms, stem cells’ regenerative medical promise is obvious. An organism develops from one or more initiating stem cells, which give rise to more specialized cell types that organize to form organs and tissues in stages that are (more or less) fixed within species. And when cells need to be replaced throughout an organism’s life, stem cells for the appropriate parts are activated. Within a multicellular organism, stem cells maintain themselves as a population via self-renewal, a pool of potentially regenerating cells for that organism’s failed, aged, or wounded parts.
So the idea of a stem cell is complex. It involves not only individual cells of various types, but also whole organisms and their parts. A stem cell, by definition, participates in processes of cell reproduction, cell development, organismal development, and organismal regeneration. A lot of biology packed into a deceptively simple term! But the real challenge is identifying cells experimentally that match that abstract concept.
What counts as a stem cell is relative to an experimental context.
The uncertainty principle of stem cells
Stem cell research is driven primarily by experiments that improve our access to individual stem cells. Stem cell researchers will try just about any available method to do this; they are experimentally ecumenical. Many, many different kinds of experiment figure in stem cell research today. The list is continually changing, as new technologies are innovated and older methods fall out of use. But there’s a core logic of experiments that aim to identify stem cells. More precisely, these experiments aim to identify a kind or variety of stem cell. Different varieties of stem cell are distinguished by species (of their source organism), by the organ in which they are found and/or (re)generate, by their source organism’s developmental stage, by their own developmental potential, by their own reproductive potential – and other character traits, often molecular or genetic. This variety indicates another respect in which stem cells are context-dependent. There are many different kinds of stem cell not only because there are many different kinds of multicellular organism (and parts of each kind), but also because stem cells are identified under various experimental conditions. Identifying stem cells is closely tied to technical advances in cell culture, visualization, and single-cell measurements. As more methods for characterizing cells are innovated, researchers are able to distinguish among more and more different types of stem cells. Our characterizations of stem cell varieties change as technologies and experimental methods change. What counts as a (kind of) stem cell is relative to an experimental context.
To see why, consider what’s involved in identifying a single living cell as a stem cell. The latter is, by definition, a cell that both reproduces itself and turns into something else; that both persists and transforms. Although intelligible in the abstract, this is a tricky thing to concretely identify. Indeed, strictly speaking, we can’t show experimentally that any single cell has specific self-renewal and differentiation abilities. The basic experimental design for identifying a stem cell in terms of these abilities is as follows:
(1) Remove cells from an organismal source and place them in an environment to self-renew.
(2) Measure cell traits in this environment.
(3) Move cells to a new environment for differentiation.
(4) Measure cell traits in the new environment.
So we start with a particular cell from some organismal source. We think this cell might be a stem cell, and we want to show this experimentally. If it’s a stem cell, then it must be able to self-renew: to divide and produce at least one stem cell like itself. So we place our candidate stem cell into an environment where that can happen, and measure the results. What exactly do we measure? We can track (sometimes) if a cell divides. And there are lots of experimental ways to characterize an individual cell: its shape, molecules on its surface, genes and proteins that it “expresses,” and so on. We can measure a bunch of those traits for the original cell – whatever traits we think might be important, which involves some guesswork – and then do the same for offspring cells later on. So we can compare across cell generations, and say (for the traits we’ve chosen) if the original and offspring cells are the same or not. If they are the same, then the original cell has self-renewed. But is it a stem cell? That cell is gone, so we can’t test it for differentiation ability. Only half the definition has been satisfied.
There’s no way to directly experimentally show that one particular cell can both self-renew and differentiate to make other kinds of cell. And there’s no way to directly experimentally show that one cell can follow more than one developmental pathway.
Suppose we test for differentiation ability first. We still measure that first cell, for a whole set of traits, and then place it in an environment where it can differentiate. But there’s no single “differentiation environment” for cells. We have to choose one pathway or another, and make an environment that encourages, say, muscle cells, or neurons, or skin, blood, bone, liver… And then we can see if cells of that type develop from the original cell. The results again are a comparison of traits, this time looking for a match between developmental products of the original cell and mature cell types of the source organism. But of course we can only try one environment, if we’re testing one particular cell. We can answer a question about one developmental pathway, at most. So we can’t discover whether that cell could self-renew, or develop along some other pathway. There’s no way to directly experimentally show that one particular cell can both self-renew and differentiate to make other kinds of cell. And there’s no way to directly experimentally show that one cell can follow more than one developmental pathway.
Claims that any single cell is a stem cell are inevitably uncertain. We can’t directly measure a single cell’s capacity for self-renewal or differentiation, separately or together.
What we need, to show that a particular cell is a stem cell, are copies of that cell. And not just genetic copies – (nearly) every cell of a multicellular organism is that. We need cells that are reproductive and developmental copies of one another; cells that perform those processes in exactly the same way, if put in the same environment. But the cells we’re looking for are stem cells – transformative by definition. Stem cells aren’t likely to “hold still” while we assemble a collection of identical copies to run experiments to test their self-renewal and differentiation abilities. We need those copies, but we can’t be sure that we have them.
Consequently, claims that any single cell is a stem cell are inevitably uncertain. We can’t directly measure a single cell’s capacity for self-renewal or differentiation, separately or together. To get those results, multiple copies of a candidate stem cell are needed. But when we do an experiment like this, we can never be sure that the cells we’re testing are exact reproductive and developmental copies of one another. That they are is a working hypothesis. That hypothesis is tied to the details of the experiment: the cell traits we measure, the environments used to detect self-renewal and differentiation. As new cell traits are discovered and made accessible to measurement, and new environments are engineered or found, the working hypothesis must be continually reassessed and revised. This means that what counts as a stem cell is tied to technologies and experimental methods for culturing, visualizing, and manipulating cells. More simply: stem cells are identified relative to particular experimental methods. Our knowledge about them accumulates by multiplying experimental contexts and relating their outcomes to one another.
Our research practices and objects of study mirror one another. Stem cells as biological entities are identified provisionally, relative to particular experimental methods and assumptions.
Living things are not isolated entities
It’s an idealization to think of stem cell experiments in isolation. In practice, stem cells are identified provisionally via lengthy sequences of experiment, guided by insights from “neighboring” experiments on other kinds of stem cell. Both stem cells and stem cell science are richly context-dependent. In this way, our research practices and objects of study mirror one another. Stem cells as biological entities are identified provisionally, relative to particular experimental methods and assumptions. Knowledge about them is updated and re-assessed as experimental methods change. These layers of context-dependence explain why stem cells are so diverse, why the scientific language used to describe them is so jargon-heavy and rapidly-changing, and why debates about cell differentiation abilities are hard to settle. The field is complex, and there’s no simple way out. Stem cell research as a whole makes progress by multiplying experimental contexts and relating their outcomes to one another. This view of science mirrors a view of life that’s increasingly popular among philosophers of biology: living things as interconnected, enmeshed in dynamically interacting contexts rather than as isolated entities.
Join the conversation