Stem Cells and Cellular Potency

A stem cell is any cell that satisfies two criteria simultaneously: it can divide to produce more stem cells indefinitely (self-renewal), and its progeny can differentiate into one or more specialized cell types. Neither property alone defines a stem cell — a cancer cell self-renews without proper differentiation, and a fully differentiated neuron cannot divide at all. The combination is what matters, and it is tested experimentally, never assumed from appearance alone.

Potency describes the breadth of that differentiation capacity, and it forms a strict hierarchy established during development:

Each step down this ladder is a one-way developmental decision under normal physiology. A cell does not spontaneously regain potency it has lost — differentiation is thermodynamically and epigenetically stable, maintained by DNA methylation, histone modifications, and self-reinforcing transcription factor networks that were established during development (see Meiosis Versus Mitosis for how the germline itself is set aside as a special potency lineage, kept totipotent-capable across generations while somatic cells are progressively restricted). The experiments described below show, however, that this restriction — while stable — is not irreversible in principle: it can be forced backward experimentally, which proves that potency loss is a regulatory state (gene expression pattern) rather than an irreversible change in the genome itself.

Step 1 — Establishing totipotency and its loss. The zygote and its immediate blastomeres (up to roughly the 8-cell stage in humans, though the exact cutoff is debated and may extend further, and is species-variable) are totipotent: each cell, isolated, can form a complete organism including placenta. As cleavage continues, cells begin to differ in position — those on the outside of the morula experience different cell-cell contact and Hippo-pathway signaling than those on the inside. Outer cells activate Hippo signaling that keeps YAP/TAZ transcriptional co-activators out of the nucleus, driving trophectoderm (placental) fate; inner cells retain nuclear YAP/TAZ activity in a way that, combined with position-dependent signaling, biases them toward inner cell mass fate. This is the first potency-narrowing decision: totipotent cells become either extraembryonic (trophectoderm) or pluripotent (inner cell mass), which can no longer make placenta.

Step 2 — Pluripotency is maintained by a core transcriptional circuit. Inner cell mass cells, and the embryonic stem (ES) cell lines derived from them, are held in the pluripotent state by three master transcription factors — Oct4 (Pou5f1), Sox2, and Nanog — that bind cooperatively at each other's enhancers and at thousands of target genes, forming an interlocking auto-regulatory loop. This circuit does two jobs at once: it activates genes needed to maintain proliferative, undifferentiated identity, and it represses differentiation genes for all three germ layers simultaneously, keeping all lineage options open. Loss of Oct4 causes ES cells to differentiate into trophectoderm-like cells; loss of Nanog causes differentiation toward primitive endoderm — showing that pluripotency is an actively maintained state, not a default one.

Step 3 — Progressive lineage restriction. As gastrulation proceeds, cells receive combinatorial signals (Nodal/Activin, BMP, Wnt, FGF gradients) that trigger commitment to ectoderm, mesoderm, or endoderm. Commitment operates by silencing pluripotency genes (Oct4, Nanog are shut off via DNA methylation and Polycomb-mediated H3K27me3 repression) while activating lineage-specific master regulators. This creates the multipotent, then oligopotent, then unipotent tiers as development proceeds — for example, a mesodermal multipotent hematopoietic stem cell (HSC) gives rise to oligopotent common myeloid and common lymphoid progenitors, which in turn generate unipotent precursors committed to a single blood cell type.

Step 4 — The self-renewal problem: how do adult stem cells avoid running out? Every time a stem cell divides, its population is at risk of either exhausting itself (if both daughters differentiate) or growing tumorously (if both daughters remain stem cells indefinitely without corresponding tissue need). Two mechanisms solve this, often used in combination by the same tissue:

• Asymmetric cell division (intrinsic mechanism): the mitotic spindle is oriented and cell-fate determinants are partitioned unequally before cytokinesis splits the cell, so the two daughters inherit different molecular contents despite identical genomes. In Drosophila neuroblasts, the protein Numb (a Notch inhibitor) localizes to one cortical crescent via the Par3/Par6/aPKC polarity complex, and the mitotic spindle aligns along this same axis, so cytokinesis (the same actomyosin contractile-ring process covered under Cytokinesis in Animal and Plant Cells) delivers Numb to only one daughter. That daughter has low Notch signaling and differentiates; the other retains Notch activity and self-renews as a neuroblast.
• Niche-dependent asymmetry (extrinsic mechanism): the stem cell divides with its mitotic spindle oriented so that only one daughter remains in physical contact with a signaling niche (a specialized microenvironment of supporting cells and matrix). The niche daughter receives self-renewal signals (e.g., Wnt, Notch ligands presented by niche cells) and stays a stem cell; the displaced daughter loses niche contact and differentiates. This is seen in Drosophila germline stem cells anchored to hub cells, and in mammalian intestinal crypt stem cells anchored near Paneth cells that supply Wnt3a and EGF.

Populations can also achieve renewal statistically rather than deterministically: a stem cell may divide symmetrically into two stem cells or two differentiated cells with some probability, and tissue homeostasis emerges from the population-level balance of these outcomes rather than every single division being asymmetric — this is well documented in intestinal crypts by lineage-tracing studies.

Step 5 — Reversal: reprogramming. Because lineage commitment is maintained by transcription factor networks and chromatin marks rather than genome changes, forcing expression of the right transcription factors can erase this memory. Yamanaka's four factors — Oct4, Sox2, Klf4, c-Myc (often abbreviated OSKM) — when overexpressed in a differentiated adult cell (e.g., a skin fibroblast), reactivate the endogenous pluripotency circuit (Oct4-Sox2-Nanog), erase repressive chromatin marks at pluripotency loci, and silence the somatic gene expression program, converting the cell into an induced pluripotent stem cell (iPSC) functionally equivalent to an ES cell.

Till and McCulloch, spleen colony-forming assay (1961–1963). Ernest McCulloch and James Till injected known numbers of mouse bone marrow cells into lethally irradiated mice (whose own marrow was destroyed and could not spontaneously regenerate). Roughly 10–14 days later, discrete nodules appeared on the spleen surface, each containing millions of cells. Three key observations followed: 1. The number of spleen colonies was linearly proportional to the number of marrow cells injected, implying each colony arose from a single progenitor cell, not from cooperative action of many cells. 2. Colonies contained multiple blood cell lineages — erythroid, myeloid, megakaryocytic — from a single founder, proving the founder cell was multipotent, not restricted to one blood lineage. 3. Cells taken from one spleen colony and re-injected into a second irradiated mouse could form new secondary colonies, again containing multiple lineages — proving the founder cell self-renewed, since it regenerated a population capable of repeating the whole process. This 1961–1963 body of work (Till & McCulloch, and with Andy Becker and Lou Siminovitch establishing clonality via chromosome-marking experiments) is generally credited as the founding demonstration of a mammalian stem cell — defined for the first time by rigorous functional criteria (self-renewal + multilineage differentiation) rather than by appearance under a microscope.

Gurdon, nuclear transfer in Xenopus (1958–1962). John Gurdon tested whether differentiation permanently alters the genome by destroying it or silencing genes irreversibly. He took nuclei from differentiated intestinal epithelial cells of tadpoles, using a UV-irradiated (nucleus-destroyed) frog egg as recipient cytoplasm, and transplanted the somatic nucleus into the enucleated egg. A fraction of these eggs developed into normal, fertile adult frogs. Since a fully differentiated cell's nucleus, placed in an egg's cytoplasm, could direct development of an entire organism, the differentiated genome must retain all the genetic information needed for totipotency — differentiation is a change in gene expression, not gene content or irreversible gene loss. This result (extended by Wilmut and colleagues to mammals with Dolly the sheep in 1996) is the conceptual foundation that made iPSC reprogramming plausible.

Yamanaka and Takahashi, induced pluripotent stem cells (2006). Shinya Yamanaka's group screened 24 candidate transcription factors known to be important in ES cell identity, retrovirally transducing them into mouse embryonic and adult fibroblasts and testing for reactivation of a pluripotency-linked reporter gene (Fbx15). By systematically removing factors from the pool of 24 and testing which reduced subsets still worked, they narrowed the requirement to just four factors: Oct4, Sox2, Klf4, and c-Myc. Fibroblasts expressing these four factors formed colonies resembling ES cells, expressed endogenous pluripotency genes, and — in later, more stringent tests — contributed to all three germ layers in chimeric mice and even to the germline. This proved that pluripotency could be induced in a fully differentiated somatic cell by a defined, minimal set of transcription factors, without needing an egg cytoplasm at all. Yamanaka shared the 2012 Nobel Prize in Physiology or Medicine with Gurdon for these complementary discoveries.

Example 1 — Interpreting a spleen colony dilution experiment. A researcher injects irradiated mice with bone marrow cell doses of 10⁴, 2×10⁴, and 4×10⁴ cells, and counts spleen colonies at day 12:

The colony count scales almost linearly with cell number (ratio stays close to ~8 colonies per 10⁴ cells across a four-fold range in dose). This linearity is the statistical signature Till and McCulloch relied on: if colonies required cooperative interaction between multiple injected cells, colony number would rise faster than linearly with dose (a higher-order/sigmoidal relationship, since the chance of the right combination of cooperating cells meeting would increase disproportionately as cell density rises). Because the relationship is linear, each colony can be attributed to a single clonogenic cell — one colony-forming unit (CFU-S) per approximately 1,250 marrow cells here (10,000/8). If a student is then told that a given colony, upon histological sectioning, contains erythrocytes, granulocytes, and megakaryocytes, they should conclude the founder cell was multipotent, and if cells from that colony can generate secondary colonies in a second irradiated host, the founder cell was also demonstrated to be self-renewing — together satisfying the operational definition of a hematopoietic stem cell.

Example 2 — Reasoning through a clinical bone marrow transplant scenario. A 34-year-old patient with acute myeloid leukemia receives a myeloablative conditioning regimen (high-dose chemotherapy plus total body irradiation) that destroys essentially all of their own hematopoietic stem cells and leukemic cells, followed by an infusion of 3 × 10⁶ CD34⁺ donor hematopoietic stem cells per kilogram of body weight (a standard clinical dose). The patient weighs 70 kg, so the total infused dose is: 3 × 10⁶ × 70 = 2.1 × 10⁸ CD34⁺ cells. Within about 2–4 weeks, the neutrophil count begins to recover, and by day 100 essentially all blood cell lineages — erythrocytes, platelets, granulocytes, lymphocytes — are of donor origin, confirmed by donor-specific chimerism testing (e.g., short tandem repeat analysis showing donor-pattern alleles in peripheral blood). Why does a relatively small number of infused cells (2.1 × 10⁸, versus roughly 3 × 10¹¹ total blood cells turned over per day in an adult) suffice to reconstitute the entire hematopoietic system for the rest of the patient's life? Because true HSCs are rare within the CD34⁺ fraction (perhaps 1 in 10³–10⁴ of CD34⁺ cells is a long-term repopulating HSC) but each one is both multipotent and capable of essentially unlimited self-renewal — exactly the property Till and McCulloch demonstrated with secondary spleen colonies. A single successfully engrafted HSC can, over years, repopulate an entire lineage tree because each division can regenerate another HSC alongside a committed progenitor, exactly as the intrinsic/niche asymmetric division mechanisms above describe. This is also why graft failure is catastrophic: if the transplanted population lacks sufficient true stem cells (as opposed to short-lived progenitors that can produce mature cells only transiently), blood counts recover initially but then collapse weeks to months later as the transient progenitors are exhausted without a self-renewing source behind them.

Shinya Yamanaka trained first as an orthopedic surgeon in Japan before turning to basic research, motivated in part by frustration at how little surgery could do for degenerative diseases. Working at Kyoto University, he took what many considered a quixotic approach: rather than trying to reverse-engineer development gene by gene using known signaling pathway logic, he ran a brute-force functional screen of candidate "stemness" transcription factors in fibroblasts, using a resistance/reporter selection system to find the minimal combination sufficient to reactivate pluripotency. The narrowing from 24 candidates to a defined four-factor cocktail — Oct4, Sox2, Klf4, and c-Myc — was itself a methodological triumph in combinatorial experimental design, not just a biological discovery.

Hematopoietic stem cell transplantation is standard of care for leukemias, lymphomas, and some inherited blood disorders (e.g., sickle cell disease, thalassemia); it works precisely because HSCs are multipotent and self-renewing, so a small graft can reconstitute an entire blood and immune system for life, as reasoned through above.

Corneal and skin stem cell grafts use limbal epithelial stem cells and epidermal stem cells (multipotent/unipotent within their tissues) to treat severe burns and corneal blindness, expanding a patient's own or a donor's stem cells ex vivo before grafting.

iPSC-derived disease modeling and drug screening: reprogramming a patient's skin or blood cells into iPSCs, then differentiating them into disease-relevant cell types (cardiomyocytes, neurons, pancreatic beta cells), lets researchers study genetic diseases in a dish using cells that carry the patient's own mutations, and screen drug candidates without invasive biopsies of inaccessible tissue like brain or heart.

Cell replacement therapy trials use iPSC- or ES-cell-derived dopaminergic neurons for Parkinson's disease and retinal pigment epithelium for macular degeneration, directly exploiting controlled differentiation of pluripotent cells down defined lineage pathways.

Cancer biology draws directly on stem cell concepts through the cancer stem cell hypothesis, which proposes that some tumors are maintained by a rare self-renewing subpopulation resistant to chemotherapy — directly extending the self-renewal/differentiation logic from normal tissue stem cells (and connecting back to how The Cell Cycle and Cancer treats uncontrolled proliferation) to explain tumor relapse after apparently successful treatment.

Agricultural and conservation biotechnology apply related somatic cell nuclear transfer techniques (descended directly from Gurdon's experiments) for livestock cloning and, more speculatively, for attempts to rescue genetic diversity in endangered species by banking and reprogramming somatic cells from individuals that can no longer breed.

Ask the tutor: How does asymmetric spindle orientation actually get coupled mechanically to the cortical polarity complex during mitosis in a stem cell?