Stem cells: what a cell can still become, and what fixes it

What makes a cell a stem cell

Almost every cell in your body does one job and will never do another. A stem cell is one of the exceptions. It is unspecialised, and it keeps two abilities that a specialised cell has given up: it can divide by mitosis to produce more cells like itself, which is self-renewal, and it can differentiate into at least one specialised type.

Both halves are needed, and questions are written to separate them. A cell that could differentiate but not renew itself would be used up after one division and the tissue would have no way of replacing it. A cell that divided without limit and never differentiated would build a mass rather than a tissue, which is the connection between this topic and the previous lesson.

These cells are not laboratory curiosities. The epithelium lining your small intestine is replaced about every five days, and roughly two million erythrocytes are produced every second. None of that replacement comes from mature cells dividing, because a mature erythrocyte has no nucleus at all and a neurone has left the cell cycle for good. It comes from stem cells in the tissue, dividing now.

Stem cell

An unspecialised cell that can divide repeatedly to produce more cells like itself, and can differentiate into one or more specialised types.

Self-renewal

Division that produces at least one daughter cell identical to the parent stem cell, so the population of stem cells is maintained.

Differentiation

The process by which a cell becomes specialised, through transcription of some of its genes and not others.

Potency

The range of cell types a stem cell is still able to become.

Four levels, and how far each one goes

Potency is examined by name, and the four names are worth learning with an example attached to each, because a question that gives you a cell and asks for its potency is really asking whether you know what that cell can and cannot make.

A totipotent cell can become every cell type of the organism and also the extra-embryonic tissues, which means the placenta and the membranes around the embryo. In a mammal that is the zygote and the cells of the first few divisions, and it does not last: within a few days no cell of a human embryo is totipotent any longer. Plants behave differently, and a single parenchyma cell from a carrot root can be grown into a whole carrot plant, which is why plant tissue culture works at all.

A pluripotent cell can become any cell type found in the body but not the extra-embryonic tissues, so it cannot by itself give rise to a whole organism. The standard example is a cell of the inner cell mass of a blastocyst, and an embryonic stem cell is exactly that cell grown in culture.

A multipotent cell can become a limited number of related types, usually the types belonging to one tissue. Quote the haematopoietic stem cells of bone marrow: they give rise to erythrocytes, to every kind of white blood cell, and to the megakaryocytes that shed platelets, and to nothing outside that family.

A unipotent cell can become one type only. The example exam papers use is the precursor that makes new heart muscle cells, though how much the adult heart renews itself, and from what, is still argued over in the research literature. The committed red cell precursor in the figure shows the same narrowing, with one correction attached: strictly it is a committed progenitor rather than a stem cell, dividing a limited number of times on its way to becoming red blood cells rather than maintaining its own population indefinitely.

One warning about the word totipotent, because it is the level candidates reach for by default. If the cell named in a question is anything other than a zygote or one of the handful of cells immediately descended from it, the answer is not totipotent.

Where stem cells are obtained from

Four sources appear in specifications, and each one comes with a potency, a practical difficulty and, in one case, an argument. Learn them as a set rather than one at a time, because comparison questions are common.

Embryonic stem cells are taken from the inner cell mass of a blastocyst about five days after fertilisation, in practice one produced by in vitro fertilisation and not returned to the uterus. They are pluripotent and they divide in culture apparently without limit, which makes them the most useful cells on this list and the most argued about: the blastocyst does not survive the procedure.

Umbilical cord blood is collected from the cord and placenta after a birth, which puts nobody at risk, and can be frozen and stored for years. The cells are multipotent and mainly haematopoietic. The limitation is quantity: one cord yields a small volume, which may be too few cells to treat an adult.

Adult tissue holds small populations of multipotent cells in bone marrow, the lining of the gut, the skin, the cornea and the brain. Bone marrow is aspirated from the pelvis under anaesthetic, so the donor takes a small risk. These cells are hard to find and to separate from the tissue around them, and their range is narrow, but cells taken from a patient and returned to the same patient carry no antigens the patient's immune system will react against.

Induced pluripotent stem cells, usually written iPS cells, are the newest of the four. An ordinary body cell such as a skin fibroblast is given four genes coding for transcription factors, and those factors switch the cell's own genes back to the pattern of a pluripotent cell. The result is pluripotent, patient-specific and made without an embryo, which is why the technique matters so much to the argument in the last section. It is still under investigation: reprogramming works on a small fraction of the cells treated, and early methods used a virus to carry the genes in, which raised the risk of tumours forming afterwards.

Nothing is lost: potency narrows by switching genes off

This is the mechanism the rest of the topic rests on, and it is the point students most often get backwards. Differentiation removes nothing. Almost every nucleated cell in your body carries the same complete genome, and what differs between a neurone and a lymphocyte is which genes are transcribed. As a cell differentiates, whole programmes of genes are shut down and the regulatory state that would reopen them is dismantled, so its potency narrows while its DNA stays exactly as it was.

Two pieces of evidence are usually quoted. In 1996 a nucleus taken from an adult sheep's mammary gland cell was transferred into an egg cell whose own nucleus had been removed, and the result was a whole sheep. That nucleus had been sitting in a specialised cell, and it still held everything needed to build a lamb. Then in 2006 came reprogramming: four transcription factors were enough to return a mouse fibroblast to a pluripotent state, and human cells followed the year after. The four genes delivered are extra copies of genes the fibroblast already carried, there to reawaken the cell's own copies, which is why the figure can truthfully show the same gene list before and after.

So potency is a property of gene regulation rather than of gene content, and that is why it can be reversed. It also explains the shape of the nested figure: each step inwards locks away the expression programmes of the options given up. What matters is which programmes remain openable rather than a simple count of genes switched off.

What they are used for, and what is only proposed

Three uses are established medicine, and it helps to say so plainly, because the topic is usually presented as being all about the future. A bone marrow transplant replaces a patient's haematopoietic stem cells and has been used since the 1960s for leukaemia and for inherited disorders of the blood and the immune system. Skin grafts for severe burns can be grown from a patient's own epidermal stem cells. Limbal stem cell grafts restore a cornea whose own stem cell population has been damaged, and restore sight with it.

Everything else on the usual list is a trial or a proposal. Insulin secreting cells for type 1 diabetes, dopamine secreting neurones for Parkinson's disease, retinal pigment cells for macular degeneration, repair after a spinal injury and repair of heart muscle after an infarction are all being worked on, and some have reached trials in patients. Saying that a treatment exists when a trial is running is the commonest overstatement in answers to this topic.

There is also a use with no transplant in it. Cells carrying a genetic disorder can be reprogrammed, grown into the affected tissue, and used to study the disorder and to test drugs on human tissue rather than on an animal.

Three difficulties are worth naming because mark schemes credit them: getting the cells to differentiate into the wanted type and no other, the risk that any undifferentiated cell left in the transplant divides into a tumour, and rejection by the immune system unless the cells came from the patient. Cost and the length of a trial belong in the same list.

The argument about embryonic sources

Questions on this carry the command words evaluate and discuss, and the marks go to the reasoning on each side and to a conclusion that follows from it. What follows sets out the positions people actually hold. It does not settle between them, and neither should an exam answer that has not argued for its conclusion.

One position is that a human embryo is a human individual from fertilisation onwards, and that its development may not be ended for the benefit of somebody else however great that benefit. On this view the number of cells and the absence of a nervous system change nothing, because what matters is what the embryo is and what it will become.

A second position is that moral status is acquired gradually. A blastocyst is about a hundred cells with no nerve tissue and no capacity to feel anything, and even in normal reproduction a large proportion of embryos never implant. On this view its claim is real but weaker than the claim of a person living with an untreatable condition.

A third argument concerns the embryos actually used. They are typically surplus from in vitro fertilisation, produced with consent and destined otherwise to be allowed to perish, so the argument runs that research adds no loss that was not going to happen. Two replies are made to it: that consenting to one outcome does not by itself make a different use permissible, and that a demand for embryos may encourage the production of more of them than a couple needs.

A fourth position is about regulation rather than principle. In the United Kingdom research on human embryos requires a licence from the Human Fertilisation and Embryology Authority, requires the consent of the donors, and may not continue beyond fourteen days after fertilisation, which is when the primitive streak appears and the point after which twinning can no longer occur. Some hold that a regulated limit is where a line can reasonably be drawn; others hold that any line drawn at a developmental landmark is arbitrary, and countries differ on where they have drawn it.

A fifth argument is that the question is being overtaken. If iPS cells can do what embryonic cells do, then an embryonic source is no longer necessary. Against that, iPS cells have not yet been shown to be equivalent: embryonic lines are still the standard they are compared with, reprogramming carries its own risk of tumours, and the comparison itself requires embryonic cells to exist.