Directed Cell Differentiation | Process and Examples

Clinical journalGlossary

Directed Cell Differentiation | Process and Examples

15 minutes read

Evidence-informed reading from the Regeneration Center.

Page last updated

Cell differentiation is the process by which a less specialized cell becomes a more specialized one, with the shape, proteins and job of a particular cell type: a red blood cell that carries oxygen, a neuron that fires, a fat cell that stores lipid. It happens on a large scale as an embryo builds its tissues, and on a smaller scale throughout life as stem cells in the adult body replace cells that wear out. Differentiation changes which genes a cell uses, not the genes it carries. Each body cell generally contains the same DNA, but each cell type switches on a different set of genes to make the proteins its work requires.[1]

Differentiation is not the same as cell division. Division makes two daughter cells; differentiation changes what a cell is. The two often happen together, because a cell on its way from stem cell to mature cell usually passes through several stages, becoming more specialized at each step.[2] Directed differentiation is the laboratory version: researchers supply signals in a planned order to steer stem cells toward one chosen cell type.

How Cells Become Specialized

Three layers of control work together to set and hold a cell’s identity.

  • Signals from outside the cell. Factors secreted by other cells, such as growth factors, are among the signals that tell a cell which direction to take.[2]
  • Transcription factors inside the cell. Signals act by turning on transcription factors, proteins that bind DNA and switch whole sets of genes on or off. A few of them can be enough to change a cell’s identity. In a 1987 experiment, adding a single gene, later named MyoD, converted mouse fibroblasts (connective tissue cells) into stable muscle-forming myoblasts.[3] Transcription factors also compete: in blood formation, GATA-1 promotes the red cell, platelet and eosinophil lines and PU.1 promotes the myeloid line, and the two proteins bind and block each other.[4]
  • Epigenetic marks. Chemical tags on DNA and on the histone proteins it is wound around keep genes switched on or off without altering the DNA sequence. As cells specialize, genes needed later in development are first held silent by histone marks, which can be reversed, while DNA methylation gives long-term silencing; in body cells it shuts down the genes that keep embryonic cells pluripotent. Which mark, if any, makes the first decision is still unknown.[5]

Some of the first evidence that specialized cells keep the genes for other cell types came from cloning experiments in frogs, later followed by nuclear transfer in mammals: the nucleus of a specialized cell can be reprogrammed to the gene activity of a different, unrelated cell type.[6] There are a few exceptions to the rule that differentiation leaves the DNA unchanged: mature red blood cells contain no DNA, and some immune cells rearrange their DNA as they start producing antibodies.[1]

Differentiation and Potency

Potency is the range of cell types a cell can still produce, and it narrows as differentiation proceeds. The fertilized egg is totipotent. The inner cell mass of the early embryo, and the embryonic stem cells grown from it, are pluripotent and can make every cell type of the body. Adult tissues hold multipotent stem cells, which make several cell types of one tissue or family; their descendants include oligopotent and unipotent progenitor cells, which make a few cell types or one. The full scale, with examples at each level, is set out on our page on undifferentiated cells.

Terminal Differentiation, Reprogramming and Transdifferentiation

A cell that has reached the end of its line and normally stops dividing is called terminally differentiated. Neurons and heart muscle cells are the usual examples. In the human heart, measurements of carbon-14 from Cold War nuclear tests built into the DNA of heart muscle cells showed that they do renew, but slowly: about 1% are replaced each year at age 25, falling to 0.45% at age 75, and fewer than half are exchanged over a normal lifetime.[7]

For a long time differentiation was thought to run only one way. Two kinds of experiment showed otherwise.

  • Reprogramming to pluripotency. In 2006, four transcription factors (Oct3/4, Sox2, c-Myc and Klf4) turned mouse fibroblasts into cells that looked and grew like embryonic stem cells, formed tumors containing tissues of all three germ layers in mice and contributed to mouse embryos. These are induced pluripotent stem cells.[8]
  • Transdifferentiation. This is the direct conversion of one specialized cell type into another without passing through a stem cell stage. Three transcription factors (Ascl1, Brn2 and Myt1l) converted mouse fibroblasts in the laboratory into neurons that made action potentials and formed working synapses.[9] In laboratory experiments with mouse cells, differentiated B cells and committed T-cell precursors were converted into working macrophages by adding the transcription factor C/EBPα.[4][10][11]

Worked Example: Blood Formation

Blood is the best-studied differentiation system. Hematopoietic (blood-forming) stem cells live in small numbers in the bone marrow and supply every blood cell type for life.[4] Their descendants narrow step by step into myeloid progenitors, which lead to red cells, platelets and most white cells, and lymphoid progenitors, which lead to lymphocytes. The oligopotent stages of this tree, and how single-cell studies have redrawn it, are covered on our oligopotent cells page. The tidy branching tree in textbooks oversimplifies how progenitors arise.[4] In medicine, a hematopoietic cell transplant works because the transplanted stem and progenitor cells carry out this whole program in the recipient.

Worked Example: Mesenchymal Stromal Cells to Bone, Cartilage and Fat

In 1999, researchers isolated mesenchymal stem cells from human bone marrow aspirates and showed that, under different culture conditions, they could be induced to become fat cells, cartilage cells or bone cells. Individual cells expanded into colonies kept all three abilities, so a single cell could hold all three potentials.[12]

That three-way test became part of the definition. The minimum criteria of the International Society for Cellular Therapy (now the International Society for Cell & Gene Therapy) require that human mesenchymal stromal cells stick to plastic in standard culture, carry the surface markers CD105, CD73 and CD90, lack CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA-DR, and differentiate into osteoblasts, adipocytes and chondroblasts in the dish.[13] The trilineage test is carried out in culture. It shows what the cells can do in a dish, not what they become after injection into a person.

Directed Differentiation in the Laboratory

Directed differentiation protocols copy the order of signals an embryo uses. Each stage gets the growth factors or small molecules that, in development, push cells to the next step, and the cells are checked for the markers expected at that stage before the next signal is added.

  • Dopamine neurons for Parkinson’s disease. Earlier dopamine neurons made from human pluripotent stem cells survived poorly after transplantation. A 2011 protocol first made midbrain floor-plate precursors, by exposing pluripotent cells to small molecules that activate the sonic hedgehog and WNT pathways, and obtained engraftable dopamine neurons by day 25. Grafted into mice, rats and monkeys with Parkinson-like damage, the cells survived and worked without overgrowth, which suggested that the earlier failures came from incompletely specified cells.[14] In 2025 two early human trials were reported. In an open-label phase I trial, 12 patients received dopamine neuron progenitors made from embryonic stem cells; there were no adverse events related to the cells at one year, and brain scans at 18 months showed graft survival.[15] In a phase I/II trial in Kyoto, seven patients received dopamine progenitors made from induced pluripotent stem cells; there were no serious adverse events, and four of the six patients assessed improved on the main motor score when off medication. Neither trial had a control group.[16]
  • Insulin-producing beta cells for type 1 diabetes. A 2014 protocol produced hundreds of millions of glucose-responsive beta cells from human pluripotent stem cells; they released insulin in response to repeated glucose challenges and reduced high blood sugar in diabetic mice.[17] In an interim, uncontrolled analysis of a phase 1-2 study of fully differentiated islets made from stem cells (zimislecel), infused into the liver’s portal vein with immunosuppression, 10 of the 12 people given the full dose no longer needed insulin injections at one year. Two of the 14 participants died, one of cryptococcal meningitis and one of progression of an existing dementia.[18]

As of November 2026, neither product is on the US Food and Drug Administration list of approved cellular and gene therapy products.[19] The list’s one islet cell product, donislecel, is made from the pancreatic cells of deceased donors.[20] In Japan, the dopamine progenitor product from the Kyoto trial (raguneprocel) received conditional, time-limited approval in 2026 on the basis of that seven-patient study. The regulator judged it likely to have some efficacy but said the evidence was limited; the approval lasts seven years, and every treated patient must be followed until the company applies for standard approval.[21]

Deterministic and Stochastic Fate Decisions

In a deterministic model, a cell’s fate follows from the signals it receives: the same inputs give the same result. In a stochastic model, chance plays a part, so identical cells in identical conditions can make different choices, while the proportions across the whole population stay predictable. Researchers distinguish two sources of this randomness. Intrinsic noise is the chance built into the chemistry of gene expression itself; extrinsic noise comes from fluctuations in other cell components, which differ from cell to cell. Experiments in bacteria showed that both contribute substantially, and that the small number of copies of some molecules inside a cell limits how precisely genes can be regulated.[22] Several systems show the stochastic side clearly.

  • Hes1 in embryonic stem cells. In mouse embryonic stem cells, expression of the Hes1 gene and the genes it controls rises and falls in cycles. Cells caught at low Hes1 tended to become neural cells and those at high Hes1 tended to become mesoderm; cells without Hes1 became neural cells more uniformly and earlier.[23] A computer model showed that intrinsic noise in the Hes1 gene network can explain this cell-to-cell variation.[24]
  • B lymphocytes. When the times at which thousands of activated mouse B cells switched antibody class, became antibody-secreting plasmablasts, divided or died were measured, each outcome behaved as an independent random process, and competition between them set the share of cells taking each path.[25] Other work in the same period found B cell fate to be directed by signals or inherited from the parent cell, depending on the circumstances.[26]
  • Bone-resorbing osteoclasts. In a cloned macrophage-like cell line (C7) grown in the laboratory and given M-CSF and RANKL, the two factors that induce osteoclasts, fewer than 10% of cells became osteoclasts, and that fraction stayed the same even after repeated cloning. No treatment tested raised it, and the authors concluded that commitment ran on “autopilot”.[27]
  • Population asymmetry in the gut. A stem cell can keep stem cell numbers steady by dividing asymmetrically, into one stem cell and one cell that goes on to differentiate. Stem cell numbers can also be balanced across a population: one stem cell divides into two differentiating cells while another divides into two stem cells. Lineage tracing in the mouse small intestine found that most divisions of Lgr5 stem cells are symmetric and that each daughter becomes a stem cell or a differentiating cell by chance.[28] As a result, the loss of one stem cell is made up by the division of a neighbor, and the clones descended from individual stem cells expand and shrink at random until one takes over the crypt or disappears, a pattern called neutral drift.[29] How single stem cells divide asymmetrically, and how self-renewal is maintained over a lifetime, is covered on our page on long-term self-renewal.

How Differentiation Is Shown

A cell’s shape alone says little about what it has become. Researchers confirm identity with marker genes and proteins measured at each stage, and then test function: dopamine neurons must release dopamine and fire, beta cells must release insulin in response to glucose, bone-forming cells must lay down mineral. The 2011 dopamine neuron work combined molecular profiling, biochemical and electrical measurements with grafts in three animal species before human trials began,[14] and the beta cell studies measured insulin release in the dish, then blood sugar in mice, then C-peptide, a marker of the body’s own insulin production, in patients.[17][18] The same standard applies to products offered to patients. For mesenchymal stromal cells, the International Society for Cell & Gene Therapy recommends calling the cells “stem cells” only when stem cell properties have been shown in the laboratory and in living animals, and backing each product with functional assays chosen for the way it is meant to work.[30] A claim that a product turns into cartilage, nerve or heart muscle in the body needs that evidence behind it.

Published Clinical Citations

[1] ^ Clark MA, Douglas M, Choi J. Biology 2e. Chapter 16 introduction: Gene Expression. OpenStax, Rice University; 2018. https://openstax.org/books/biology-2e/pages/16-introduction

[2] ^ National Institutes of Health. Stem Cell Basics. NIH Stem Cell Information. https://stemcells.nih.gov/info/basics/stc-basics

[3] ^ Davis RL, Weintraub H, Lassar AB. Expression of a single transfected cDNA converts fibroblasts to myoblasts. Cell. 1987;51(6):987-1000. https://doi.org/10.1016/0092-8674(87)90585-x

[4] ^ Orkin SH, Zon LI. Hematopoiesis: an evolving paradigm for stem cell biology. Cell. 2008;132(4):631-644. https://doi.org/10.1016/j.cell.2008.01.025

[5] ^ Reik W. Stability and flexibility of epigenetic gene regulation in mammalian development. Nature. 2007;447(7143):425-432. https://doi.org/10.1038/nature05918

[6] ^ Gurdon JB, Melton DA. Nuclear reprogramming in cells. Science. 2008;322(5909):1811-1815. https://doi.org/10.1126/science.1160810

[7] ^ Bergmann O, Bhardwaj RD, Bernard S, et al. Evidence for cardiomyocyte renewal in humans. Science. 2009;324(5923):98-102. https://doi.org/10.1126/science.1164680

[8] ^ Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell. 2006;126(4):663-676. https://doi.org/10.1016/j.cell.2006.07.024

[9] ^ Vierbuchen T, Ostermeier A, Pang ZP, et al. Direct conversion of fibroblasts to functional neurons by defined factors. Nature. 2010;463(7284):1035-1041. https://doi.org/10.1038/nature08797

[10] ^ Xie H, Ye M, Feng R, Graf T. Stepwise reprogramming of B cells into macrophages. Cell. 2004;117(5):663-676. https://doi.org/10.1016/s0092-8674(04)00419-2

[11] ^ Laiosa CV, Stadtfeld M, Xie H, et al. Reprogramming of committed T cell progenitors to macrophages and dendritic cells by C/EBPα and PU.1 transcription factors. Immunity. 2006;25(5):731-744. https://doi.org/10.1016/j.immuni.2006.09.011

[12] ^ Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult human mesenchymal stem cells. Science. 1999;284(5411):143-147. https://doi.org/10.1126/science.284.5411.143

[13] ^ Dominici M, Le Blanc K, Mueller I, et al. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy. 2006;8(4):315-317. https://doi.org/10.1080/14653240600855905

[14] ^ Kriks S, Shim JW, Piao J, et al. Dopamine neurons derived from human ES cells efficiently engraft in animal models of Parkinson’s disease. Nature. 2011;480(7378):547-551. https://doi.org/10.1038/nature10648

[15] ^ Tabar V, Sarva H, Lozano AM, et al. Phase I trial of hES cell-derived dopaminergic neurons for Parkinson’s disease. Nature. 2025;641(8064):978-983. https://doi.org/10.1038/s41586-025-08845-y

[16] ^ Sawamoto N, Doi D, Nakanishi E, et al. Phase I/II trial of iPS-cell-derived dopaminergic cells for Parkinson’s disease. Nature. 2025;641(8064):971-977. https://doi.org/10.1038/s41586-025-08700-0

[17] ^ Pagliuca FW, Millman JR, Gürtler M, et al. Generation of functional human pancreatic β cells in vitro. Cell. 2014;159(2):428-439. https://doi.org/10.1016/j.cell.2014.09.040

[18] ^ Reichman TW, Markmann JF, Odorico J, et al. Stem cell-derived, fully differentiated islets for type 1 diabetes. New England Journal of Medicine. 2025;393(9):858-868. https://doi.org/10.1056/NEJMoa2506549

[19] ^ US Food and Drug Administration. Approved Cellular and Gene Therapy Products. Content current as of 30 October 2026. https://www.fda.gov/vaccines-blood-biologics/cellular-gene-therapy-products/approved-cellular-and-gene-therapy-products

[20] ^ US Food and Drug Administration. FDA approves first cellular therapy to treat patients with type 1 diabetes (Lantidra). News release, 28 June 2023. https://www.fda.gov/news-events/press-announcements/fda-approves-first-cellular-therapy-treat-patients-type-1-diabetes

[21] ^ Ministry of Health, Labour and Welfare; Pharmaceuticals and Medical Devices Agency. Report on the Deliberation Results and Review Report: AMCHEPRY (raguneprocel), Sumitomo Pharma Co., Ltd. 19 February 2026 (English translation). https://www.pmda.go.jp/files/000281617.pdf

[22] ^ Elowitz MB, Levine AJ, Siggia ED, Swain PS. Stochastic gene expression in a single cell. Science. 2002;297(5584):1183-1186. https://doi.org/10.1126/science.1070919

[23] ^ Kobayashi T, Mizuno H, Imayoshi I, et al. The cyclic gene Hes1 contributes to diverse differentiation responses of embryonic stem cells. Genes & Development. 2009;23(16):1870-1875. https://doi.org/10.1101/gad.1823109

[24] ^ Sturrock M, Hellander A, Matzavinos A, Chaplain MAJ. Spatial stochastic modelling of the Hes1 gene regulatory network: intrinsic noise can explain heterogeneity in embryonic stem cell differentiation. Journal of the Royal Society Interface. 2013;10(80):20120988. https://doi.org/10.1098/rsif.2012.0988

[25] ^ Duffy KR, Wellard CJ, Markham JF, et al. Activation-induced B cell fates are selected by intracellular stochastic competition. Science. 2012;335(6066):338-341. https://doi.org/10.1126/science.1213230

[26] ^ Tarlinton D. B-cell differentiation: instructive one day, stochastic the next. Current Biology. 2012;22(7):R235-R237. https://doi.org/10.1016/j.cub.2012.02.045

[27] ^ Hayashi S, Murata A, Okuyama K, et al. Stochastic differentiation into an osteoclast lineage from cloned macrophage-like cells. Biochemical and Biophysical Research Communications. 2012;428(2):303-308. https://doi.org/10.1016/j.bbrc.2012.10.052

[28] ^ Snippert HJ, van der Flier LG, Sato T, et al. Intestinal crypt homeostasis results from neutral competition between symmetrically dividing Lgr5 stem cells. Cell. 2010;143(1):134-144. https://doi.org/10.1016/j.cell.2010.09.016

[29] ^ Lopez-Garcia C, Klein AM, Simons BD, Winton DJ. Intestinal stem cell replacement follows a pattern of neutral drift. Science. 2010;330(6005):822-825. https://doi.org/10.1126/science.1196236

[30] ^ Viswanathan S, Shi Y, Galipeau J, et al. Mesenchymal stem versus stromal cells: International Society for Cell & Gene Therapy (ISCT) Mesenchymal Stromal Cell committee position statement on nomenclature. Cytotherapy. 2019;21(10):1019-1024. https://doi.org/10.1016/j.jcyt.2019.08.002

Continue exploring

Three related medical articles and reference pages.