Hematopoiesis: Origin of Immune Cells

Hematopoiesis (from Greek haima, blood, and poiesis, to make) is the formation of all cellular blood components. In the adult, it occurs almost exclusively in the bone marrow, principally within the flat bones (sternum, pelvis, vertebrae, ribs) and the proximal ends of long bones. During fetal development, the site shifts sequentially — yolk sac, then fetal liver and spleen, then bone marrow — a migration pattern that reflects evolving niches rather than any change in the underlying stem cell biology. At the apex of the system is the hematopoietic stem cell (HSC), defined by two properties that must both be demonstrated experimentally, not merely inferred from appearance: self-renewal (a division that produces at least one daughter identical to the parent, maintaining the stem cell pool indefinitely) and multipotency (the capacity to differentiate into every mature blood lineage). HSCs are rare — roughly 1 in 10,000 to 1 in 100,000 bone marrow cells — and mostly quiescent, cycling only occasionally to replenish the pool or respond to demand. Most of the actual cell production is carried out by their downstream descendants, transit-amplifying progenitor populations that divide rapidly but have progressively restricted potential. The central organizing concept is lineage commitment through progressive restriction: an HSC first branches into a multipotent progenitor, which then splits into the common myeloid progenitor (CMP) and common lymphoid progenitor (CLP). The CMP further branches into megakaryocyte-erythroid progenitors (making platelets and red cells) and granulocyte-monocyte progenitors (making neutrophils, eosinophils, basophils, monocytes/macrophages, and some dendritic cells). The CLP gives rise to B lymphocytes, T lymphocyte precursors (which complete maturation in the thymus), natural killer cells, and some dendritic cells. Each restriction step is driven by changes in transcription factor expression (e.g., PU.1, GATA-1, GATA-2, Ikaros) that open or close chromatin at lineage-specific gene loci, and is reinforced — not initiated — by extracellular cytokine signaling, a subtlety established by later single-cell studies that refined the earlier "instructive cytokine" model.

Step 1 — The niche. HSCs reside in specialized microenvironments within bone marrow called stem cell niches, typically adjacent to sinusoidal blood vessels (the perivascular niche) and near bone-lining osteoblasts (the endosteal niche). Niche cells secrete stem cell factor (SCF, the ligand for the c-Kit receptor) and CXCL12 (which binds CXCR4 on HSCs), signals that keep HSCs anchored, quiescent, and viable. Disrupting CXCR4-CXCL12 signaling (pharmacologically, with the drug plerixafor) mobilizes HSCs out of the marrow into peripheral blood — the basis of modern stem cell collection for transplantation.

Step 2 — Asymmetric and symmetric division. An activated HSC can divide symmetrically (both daughters remain stem cells, expanding the pool), asymmetrically (one daughter remains a stem cell, one becomes a multipotent progenitor, MPP), or symmetrically toward differentiation (both daughters differentiate, depleting the stem pool). Under homeostasis, division is infrequent — human HSCs are estimated to divide roughly once every 40 weeks on average — but the rate rises sharply after marrow injury or blood loss.

Step 3 — Progressive lineage restriction. The MPP loses self-renewal capacity but retains broad differentiation potential, then bifurcates:

• Common myeloid progenitor (CMP) → megakaryocyte-erythroid progenitor (MEP) → erythrocytes and platelets; and → granulocyte-monocyte progenitor (GMP) → neutrophils, eosinophils, basophils/mast cells, monocytes (which become macrophages or some dendritic cells in tissues).
• Common lymphoid progenitor (CLP) → pro-B cells → B lymphocytes (mature in bone marrow); pro-T cells (migrate to the thymus for maturation) → T lymphocytes; and → NK cells and some dendritic cell subsets.

Step 4 — Cytokine-driven proliferation and maturation. Each progenitor stage displays specific cytokine receptors, and locally or systemically produced cytokines drive proliferation and terminal differentiation:

Step 5 — Feedback regulation. The system is a set of nested negative-feedback loops. The clearest example is erythropoiesis: falling blood oxygen tension is sensed by renal peritubular fibroblast-like cells via the transcription factor HIF (hypoxia-inducible factor); under low oxygen, HIF-α escapes degradation, drives EPO gene transcription, and circulating EPO rises. EPO binds its receptor on erythroid progenitors, triggering JAK2-STAT5 signaling that promotes proliferation and survival (blocking apoptosis) of red cell precursors. As red cell mass and oxygen delivery rise, the hypoxic stimulus resolves and EPO production falls — a classic homeostatic loop. Neutrophil production is regulated analogously: infection triggers macrophages and endothelial cells to release G-CSF, which accelerates neutrophil output from marrow reserves within hours (the "left shift" seen in blood counts during infection) and stimulates new production over days.

Step 6 — Terminal maturation and release. Mature cells are released into blood when they reach defined checkpoints: reticulocytes (enucleated, still RNA-containing erythroid cells) enter blood and lose their RNA within 1-2 days to become mature erythrocytes; neutrophils are stored in a marrow reserve pool roughly 10-15 times the size of the circulating pool and released rapidly upon G-CSF/inflammatory signaling; naive B cells complete central tolerance checkpoints in marrow before export, while T cell precursors leave marrow uncommitted and complete positive/negative selection in the thymus.

The existence of a self-renewing, multipotent hematopoietic stem cell was not obvious from histology alone — marrow smears show a continuum of maturing cells, and nothing under a microscope proves that a single primitive cell can both renew itself and generate every lineage. That proof required a functional assay, and it came from James Till and Ernest McCulloch at the Ontario Cancer Institute in Toronto, working with graduate student Andy Becker and technician Lou Siminovitch in 1961-1963.

The spleen colony-forming assay. Till and McCulloch irradiated mice with a dose of radiation sufficient to destroy their endogenous marrow (a dose that would otherwise be lethal), then injected known numbers of donor bone marrow cells intravenously. Ten to fourteen days later, they found discrete nodules — colonies — on the surface of the recipient's spleen. Critically, the number of colonies was directly proportional to the number of marrow cells injected, implying that each colony arose from a single progenitor cell (later called a colony-forming unit-spleen, or CFU-S) rather than from cell aggregation.

Proving both self-renewal and multipotency. Two further observations turned this from a proliferation assay into proof of stemness: 1. Multipotency: When individual spleen colonies were dissected out and examined, single colonies were shown by Till, McCulloch, and colleagues (with Becker, 1963) to contain multiple differentiated lineages — erythroid, granulocytic, and megakaryocytic cells — simultaneously, all derived from what had to be a single founding cell. This directly demonstrated that one cell can give rise to more than one blood lineage. 2. Self-renewal: When cells from a single primary spleen colony were harvested and reinjected into a second irradiated mouse, they generated new secondary spleen colonies — proof that the founding cell had reproduced itself, not merely differentiated and been exhausted. Serial transplantation could be repeated across several rounds, establishing that self-renewal capacity, though finite, was a real, heritable property of individual cells.

Quantitative rigor: proving single-cell origin genetically. A remaining worry was that a colony might arise from clumps of cooperating cells rather than one true clone. Till, McCulloch, and Siminovitch settled this using a radiation-induced chromosome marker: they irradiated donor marrow cells with a low, sub-lethal dose sufficient to produce distinct, stable, visible chromosomal aberrations in a subset of cells, then transplanted them. If a colony arose from a single cell, every cell within that colony — regardless of which lineage it had differentiated into — should carry the identical unique chromosomal marker. This is exactly what they found (Becker, McCulloch & Till, Nature, 1963): all cells in a given spleen colony, whether erythroid, granulocytic, or megakaryocytic, shared one distinctive marker chromosome, proving clonal, single-cell origin. Together, these experiments — proportionality of colony number to cell dose, multilineage content of single colonies, serial transplantability, and clonal marking — constitute the founding proof of the stem cell concept in any tissue, not just blood. They established, for the first time in any system, that a single somatic cell could be both self-renewing and multipotent, a result now considered one of the cornerstones of modern stem cell biology and regenerative medicine.

Later refinements. Modern flow cytometric purification (using surface markers such as CD34, CD38, CD90/Thy-1, and lineage-negative gating) allows isolation of near-homogeneous human and mouse HSC populations, and single-cell transplantation experiments (one purified cell reconstituting an entire irradiated animal's blood system for its lifetime) have since directly confirmed Till and McCulloch's conclusions at the single-cell level, without relying on statistical colony-counting.

Example 1 — Interpreting a CFU-S dose-response experiment. A researcher injects irradiated mice with bone marrow cell suspensions at three different doses and counts spleen colonies 10 days later:

Question: is colony formation consistent with single-cell clonal origin, and what is the approximate CFU-S frequency? Reasoning: if each colony arises from one progenitor, colony number should scale linearly with cells injected. Check the ratios: 50,000/10,000 = 5-fold increase in cells, and 21/4 ≈ 5.25-fold increase in colonies — close to linear. 200,000/10,000 = 20-fold cells, and 83/4 ≈ 20.75-fold colonies — again consistent with linearity (small deviations reflect biological/counting noise). This linear relationship is exactly the signature Till and McCulloch used to argue that each colony has a single-cell origin, since aggregation-based colony formation would produce a nonlinear (supralinear) relationship. Estimating CFU-S frequency: averaging the three ratios (4/10,000, 21/50,000, 83/200,000) gives approximately 0.0004, 0.00042, 0.000415 — about 1 CFU-S per 2,400-2,500 marrow cells, in the same order of magnitude as classically reported values (roughly 1 in 10⁴-10⁵ for the more primitive true HSC, higher for the CFU-S which includes some more mature progenitors).

Example 2 — Clinical reasoning: post-chemotherapy neutropenia and G-CSF. A 54-year-old woman receives cytotoxic chemotherapy for breast cancer. Chemotherapy preferentially kills rapidly dividing cells, including marrow progenitors (GMPs), but largely spares quiescent HSCs. Her absolute neutrophil count (ANC), normally 2,500-7,500 cells/µL, falls to 300 cells/µL by day 10 (severe neutropenia, defined as ANC < 500/µL), putting her at high risk of life-threatening bacterial infection. Reasoning through the mechanism: neutrophils have a circulating half-life of only about 8-20 hours, so any interruption in new production is reflected in blood counts within one to two days — unlike red cells (lifespan ~120 days), where a similar production block takes weeks to manifest as anemia. This explains why neutropenia is the earliest and most dangerous cytopenia after chemotherapy, while anemia develops more slowly. Clinical intervention: the oncologist prescribes recombinant human G-CSF (filgrastim). Mechanistically, G-CSF binds its receptor on granulocyte-monocyte progenitors and maturing neutrophils, (1) accelerating the mitotic rate of GMPs, (2) shortening the post-mitotic maturation time in the marrow storage pool, and (3) triggering release of the marrow's neutrophil reserve pool (roughly ten times the circulating pool size) into blood. The ANC typically rises within 1-3 days of starting G-CSF, well before newly stimulated progenitors could complete a full differentiation cycle from scratch (which takes about 10-14 days) — demonstrating that acute recovery draws heavily on accelerated release of the pre-existing marrow reserve, with de novo increased production contributing over the following one to two weeks.

Bone marrow and HSC transplantation. Transplantation of HSCs (from marrow, mobilized peripheral blood, or umbilical cord blood) is curative therapy for leukemias, lymphomas, aplastic anemia, and severe combined immunodeficiency (SCID). Understanding the hematopoietic tree explains why engraftment is monitored lineage by lineage, and why graft failure or delayed lymphoid reconstitution creates distinct clinical risks (neutropenic infection early, opportunistic viral/fungal infection later during T cell reconstitution).

Recombinant cytokine therapy. Recombinant erythropoietin treats anemia of chronic kidney disease (replacing the kidney's failing EPO output); recombinant G-CSF (filgrastim, pegfilgrastim) shortens chemotherapy-induced neutropenia and is used to mobilize HSCs from marrow into peripheral blood (combined with CXCR4 antagonists like plerixafor) for stem cell collection before transplant; thrombopoietin receptor agonists (romiplostim, eltrombopag) treat chronic immune thrombocytopenia.

Leukemia as a disease of blocked differentiation. Many leukemias arise from mutations that block normal progenitor maturation at a specific branch point of the hematopoietic tree (e.g., PML-RARA fusion blocking granulocyte maturation in acute promyelocytic leukemia), causing accumulation of immature blasts. Differentiation therapy (all-trans retinoic acid in APL) works by forcing the blocked progenitors to resume maturation — a direct clinical application of hematopoietic tree logic.

Gene therapy and genome editing. Because HSCs are self-renewing and can be harvested, edited ex vivo, and reinfused, they are prime targets for gene therapy — for example, autologous HSC gene therapy corrections for sickle cell disease and beta-thalassemia (including CRISPR-based approaches reactivating fetal hemoglobin), and for adenosine deaminase-deficient SCID.

Vaccine and infection immunology, and foreshadowing phagocyte biology. The rate and lineage bias of hematopoiesis shift during infection ("emergency granulopoiesis" and "emergency myelopoiesis"), diverting stem cell output toward neutrophils and monocytes to meet acute demand — setting up the mechanisms of pathogen recognition and killing carried out by these freshly minted phagocytes, the subject taken up next in the study of Phagocytosis: Mechanism and Killing.

Ask the tutor: How does emergency granulopoiesis during infection actually reprogram hematopoietic stem cell output toward myeloid cells, and how does this connect to the phagocyte killing mechanisms covered next?