What makes a stem cell a stem cell

Under a microscope, most stem cells are a letdown. A human blood stem cell is a small round cell with a large nucleus, almost identical to the lymphocytes drifting next to it. A flask of mesenchymal stromal cells looks like any flask of fibroblasts: long spindle shapes in loose parallel streams. Even a colony of human pluripotent stem cells is just a flat, tightly packed island of small cells with a crisp edge.

That is the first thing worth knowing. "Stem cell" is not a shape or a stain. It describes behavior, and behavior has to be shown over time.

Two abilities, both required

A cell earns the name when it can do two things.

The first is self-renewal: when it divides, at least one daughter stays a stem cell with the parent's abilities. Sometimes one division yields one stem cell and one cell headed for specialization. Sometimes both daughters stay stem cells, or both move on, and the balance is kept across the population. Either way, the pool does not run dry.

The second is potency, the capacity to give rise to specialized cell types. A stem cell near the base of the small intestine produces absorptive cells, mucus-making goblet cells and several other types, replacing a lining that turns over every few days.

You need both. Progenitor cells, the busy middle tier in most tissues, multiply fast and make mature cells, but after a limited number of rounds they are spent. And many fully specialized cells, like most neurons in the adult brain, rarely divide at all.

Both properties concern the future, so they are proven over weeks or months. The gold standard for a blood stem cell is a transplant: put a candidate cell into an animal whose own blood system has been knocked out and see whether it rebuilds every blood lineage and keeps doing so. Then move cells from that animal into a second one. If the graft works again, the original cell renewed itself. Anything short of that is suggestive, not proof.

The potency ladder

Biologists sort stem cells by how many cell types they can make. The categories are tidier on paper than at the bench, but they are a useful map.

• Totipotent: the fertilized egg and the cells of the first few divisions. They can form the embryo and the tissues that support it, including the placenta.
• Pluripotent: cells that can make every cell type of the body, from all three embryonic germ layers (ectoderm, mesoderm and endoderm), plus germ cells. The inner cell mass of an early embryo is pluripotent, and so are embryonic stem cells and induced pluripotent stem cells grown in the lab. On their own they cannot build a whole organism with its placenta.
• Multipotent: cells that make several types within one family. A blood stem cell makes red cells, platelets and the many kinds of white cell, but not liver or nerve.
• Unipotent: cells that make one type while still renewing themselves. Muscle satellite cells, which sit along skeletal muscle fibers and repair them after injury, are the usual example.

In the body, cells move down this ladder and almost never back up. The surprise of recent decades is that in a dish they can be pushed back.

Where stem cells come from

Embryonic stem cells

Embryonic stem cells come from the inner cell mass of a blastocyst, an embryo a few days after fertilization, usually one created during IVF and later donated. Mouse embryonic stem cells were first grown in 1981, and James Thomson's group at the University of Wisconsin reported human lines in 1998.

A good culture is easy to recognize. The colonies are compact with sharp borders, and the cells have big nuclei, very little cytoplasm and conspicuous nucleoli. When a colony starts to differentiate, its edge loosens and the cells flatten and spread like fried eggs, which is why people who grow these lines spend so much time scraping away bad patches by hand. The ethical debate over using embryos is real, but it is a separate question from what the cells can do.

Adult (tissue) stem cells

Most organs keep a reserve of tissue stem cells, best described in bone marrow, the gut lining, skin, hair follicles and skeletal muscle. They are rare, mostly quiet, and committed to their own tissue. A gut stem cell makes gut.

A large group sits awkwardly here. Mesenchymal stromal cells, grown from marrow, fat or cord tissue, are often sold as "mesenchymal stem cells". In culture they stick to plastic, carry a standard set of surface markers and can be coaxed into bone, cartilage and fat cells, the minimal criteria published by the International Society for Cellular Therapy in 2006. Whether they self-renew and make several tissues inside the body is much less clear, so many researchers prefer "stromal". Our explainer on bone marrow and mesenchymal stromal cells covers what they are and are not.

Induced pluripotent stem cells

In 2006 Shinya Yamanaka's lab in Kyoto showed that switching on four genes in mouse skin cells could turn them back into pluripotent cells. The human version followed in 2007, and Yamanaka shared the 2012 Nobel Prize in Physiology or Medicine with John Gurdon. These induced pluripotent stem (iPS) cells grow in the same tight colonies as embryonic stem cells and can be made from one person's skin or blood.

They are not perfect copies. A line can carry traces of the cell type it came from, and reprogramming plus long culture can introduce genetic changes, so each line has to be checked. Our longer piece on induced pluripotent stem cells covers how they are made and where they are being tested.

The niche: why location matters

A stem cell does not set its own behavior. It reads the signals around it. In 1978 the British scientist Raymond Schofield proposed that blood stem cells depend on a particular local environment in the bone marrow, which he called a niche, and the idea has held up across many tissues.

A niche is the neighborhood: nearby support cells, the proteins they release, the matrix the stem cell is anchored to, local oxygen levels, even how stiff the surroundings are. Intestinal stem cells sit at the bottom of narrow pits called crypts, wedged between cells that feed them growth signals. In bone marrow, blood stem cells are found mostly close to small blood vessels.

The niche keeps most stem cells resting, then calls them into action after injury or blood loss. It is also the main reason stem cells are hard to grow outside the body. In a flask, a blood stem cell tends to differentiate within days. Much of what we describe on our page about stem cell technology, from culture media and growth factors to scaffolds, is an attempt to imitate a niche well enough to keep cells in the state you want.

Blood stem cells: the best-understood example

If you want a thoroughly proven stem cell, the blood system is the model. In the early 1960s James Till and Ernest McCulloch, working in Toronto, injected bone marrow cells into irradiated mice and found small lumps on the spleen. Each grew from a single cell and held several kinds of blood cell, and some contained cells that formed new lumps in other mice. Many lineages from one cell, plus renewal: the definition this article started with.

Clinical practice followed. Bone marrow transplantation, developed over decades by E. Donnall Thomas and others (Thomas shared the 1990 Nobel Prize), works because a modest number of donor stem cells can rebuild a patient's entire blood and immune system. In 1988 Eliane Gluckman's team in Paris performed the first cord blood transplant, for a child with Fanconi anemia. Today blood stem cells from marrow, mobilized blood or cord blood treat blood cancers, inherited blood disorders and some immune deficiencies. That is what the long cord blood disease lists really describe, as we explain in conditions treated with cord blood transplants.

Even here, the cells are hard to see. Transplant labs count human blood stem and progenitor cells by a surface protein called CD34, measured on a flow cytometer, because by eye nothing separates them from ordinary white cells. The long-lived stem cells proper are a small minority even inside the CD34-positive group. Most daily blood production comes from progenitors further down the line, while the true stem cells divide rarely and stay in reserve.

Myth vs what the evidence says

A quick test for any stem cell claim

When "stem cell" turns up in a headline or on a label, four questions sort out most of it.

1. Which cell, from which tissue? A claim that names no source is a warning sign.
2. What potency has been shown, and in what test? A cell that formed bone in a culture dish has shown what it can do in a culture dish.
3. Was self-renewal measured, or assumed?
4. Is the use approved, being tested in a registered trial, or neither? The FDA's information page on regenerative medicine therapies and the ISSCR are good places to check.

The second question does the most work. Dish assays are useful screens, but what settles the matter is the test blood stem cells passed long ago in Toronto: put the cells in, wait months, and show that the tissue they built is still being renewed after a second transplant.