Common myeloid progenitor (CMP) — the fork at which KMT2A leukaemia almost always turns one way
also called CMP · common myeloid progenitor · Lin-CD34+CD38+IL-3Ra-lo CD45RA-
The common myeloid progenitor is the bone-marrow cell that still has a choice: it can go on to make neutrophils and monocytes, or red cells and platelets. KMT2A-rearranged leukaemia almost always ends up on the first branch, and that lopsidedness is itself a clue about what the fusion can and cannot do.
The same thing, for a scientist
The CMP (CL:0000049) is the myeloid-restricted progenitor at the branch point between the GMP and the MEP, prospectively isolable in human marrow and cord blood as Lin-CD34+CD38+ with low IL-3R-alpha and negative CD45RA [E1], defined in mouse as the clonogenic progenitor giving rise to all myeloid lineages whose commitment to GM or MegE fates is mutually exclusive [E2]. It is included here to complete the progenitor tree this graph half-describes and because the asymmetry across its two branches is a real observation about KMT2A biology: `cell-type-gmp` is the transformation target in the founding LSC experiment, the destination of therapy-associated lineage switch and the most menin-sensitive state, while `cell-type-mep` contributes no cell line to the standard KMT2A panel and about 7 percent of a rare paediatric subtype. The entry carries a heavy caveat: single-cell work has repeatedly failed to find the CMP as a discrete compartment, resolving instead a continuum of low-primed progenitors from which unilineage cells emerge directly [E3].
What the CMP is, and the honest state of the evidence about it
The textbook tree runs: stem cell → multipotent progenitor → common myeloid progenitor → either granulocyte-monocyte progenitor or megakaryocyte-erythroid progenitor → mature cells. The CMP is the last cell in that scheme that can still make any myeloid cell type.
It was defined in mouse as a clonogenic progenitor giving rise to all myeloid lineages, with the further claim that commitment to the GM branch and to the MegE branch are mutually exclusive events 2, and located in human marrow by surface markers a couple of years later 1.
The caveat has to come immediately after the definition, because in this case it is not a footnote. Integrating flow cytometry, transcriptomics and function at single-cell resolution, individual human HSCs were found to acquire lineage biases gradually along multiple directions without passing through discrete hierarchically organised progenitor populations; unilineage-restricted cells emerge directly from a continuum of low-primed undifferentiated haematopoietic stem and progenitor cells 3. Separately, the hierarchy was shown to be reshaped across ontogeny: fetal liver contains many oligopotent progenitors with intermingled myeloid, erythroid and megakaryocytic fates, adult marrow contains almost none 4.
So the CMP is best treated in this graph as a gate, a reproducible sorting coordinate that captures a real region of the landscape, and not as a cell that exists as a discrete entity waiting to make a decision.
The asymmetry across the two branches is the KMT2A-relevant fact
Downstream of this gate the two branches are treated very differently by KMT2A fusions, and the contrast is stark enough to be worth stating as an observation in its own right:
| GMP branch | MEP branch | |
|---|---|---|
| founding LSC experiment | MLL-AF9 in committed GMP produces LSCs transferring leukaemia at 4 cells | not reported |
| cell lines in the standard KMT2A panel | 6 of 7 AML lines are monocytic/monoblastic | 0 of 11 rearranged lines |
| lineage-switch destination | GMP genes preferentially activated | not reported |
| menin-inhibitor sensitivity | GMP-like is the most sensitive class ex vivo | not reported |
| clinical frequency | dominant myeloid presentation | about 7 percent of paediatric AMkL |
The absence on the right is not merely an absence of study. This project's own DepMap lineage check found zero erythroid or megakaryocytic lines among 11 KMT2A-rearranged lines while eight such lines sat in the comparison group, so the asymmetry is visible in an unbiased panel and not only in the literature's attention.
Why the fusion should be so nearly excluded from one branch of a single fork is an open mechanistic question this graph does not answer. Candidate explanations — that the required chromatin context does not exist in MegE-primed cells, that the relevant HOXA program is incompatible with GATA1-driven differentiation, or simply that MegE-committed cells do not survive the transformation — are all testable and none has been tested.
| ontology | CL:0000049 |
|---|---|
| location | Bone marrow after birth; fetal liver before birth |
| status of the compartment | treat as a sorting gate on a continuum, not a discrete cell; single-cell work does not resolve it as a distinct population in adult homeostasis |
markers
[ "Lineage-", "CD34+", "CD38+", "IL3RA/CD123 low", "CD45RA-" ]
lineage potential
[ "granulocyte", "monocyte", "erythrocyte", "megakaryocyte" ]
immediate downstream
[ "cell-type-gmp", "cell-type-mep" ]
kmt2a relevance
{
"branch_asymmetry": "KMT2A fusions are heavily over-represented on the GMP branch and nearly absent from the MEP branch; 0 of 11 rearranged DepMap lines are erythroid or megakaryocytic",
"unexplained": "no published work tests why the MegE branch is nearly closed to KMT2A fusions"
}atlas refs
[
{
"label": "Prospective isolation of human CMP, GMP and MEP",
"pmid": "12193648",
"title": "Prospective isolation of human clonogenic common myeloid progenitors.",
"authors": "Manz MG, Miyamoto T, Akashi K, et al.",
"venue": "Proc Natl Acad Sci U S A",
"year": 2002,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"label": "Continuous lineage commitment from CLOUD-HSPCs",
"pmid": "28319093",
"title": "Human haematopoietic stem cell lineage commitment is a continuous process.",
"authors": "Velten L, Haas SF, Raffel S, et al.",
"venue": "Nat Cell Biol",
"year": 2017,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]- If the progenitor hierarchy is a continuum, then stratifying patients or models by inferred progenitor stage imposes boundaries the biology does not have, and any therapy assigned on that basis inherits the arbitrariness of the boundary.
- The branch asymmetry that makes KMT2A look like a GMP-branch disease is measured in adult-derived cell lines and adult-defined gates, so a therapy designed around it may be designed around the wrong hierarchy for an infant.
- Why is the megakaryocyte-erythroid branch nearly closed to KMT2A fusions? Retroviral introduction of MLL-AF9 into sorted human MEP, the direct analogue of the 2006 GMP experiment, has apparently never been reported.
- If lineage commitment is continuous rather than stepwise, is 'the cell of origin' even a well-posed question for this disease, or should the graph be asking which region of the continuum is permissive?
- Do the marker gates that define CMP in adult marrow identify anything meaningful in fetal liver, where infant KMT2A-rearranged leukaemia begins?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Human clonogenic common myeloid progenitors and their downstream granulocyte/macrophage and megakaryocyte/erythrocyte progenitors reside in the lineage-negative CD34+CD38+ fraction of adult bone marrow and cord blood and are distinguishable by expression of the IL-3 receptor alpha chain and CD45RA.
Prospective isolation of human clonogenic common myeloid progenitors · Manz MG, Miyamoto T, Akashi K, Weissman IL · Proceedings of the National Academy of Sciences USA 2002What this evidence cannot show
Marker-defined gates with in vitro clonogenic readout; the paper demonstrates that the sorted fraction is clonogenic and multipotent in bulk, which does not establish that every cell in it is. Read: abstract only.
Evidence that argues against this (1)
- Single-cell integration of flow, transcriptomic and functional data found that human HSCs acquire lineage biases gradually along multiple directions WITHOUT passing through discrete hierarchically organised progenitor populations, with unilineage-restricted cells emerging directly from a continuum of low-primed HSPCs. On that account CMP is a gate that captures a region of a continuum, not a cell type, and conclusions that depend on the CMP being a discrete decision point do not follow. Read: abstract only. Velten L et al. 2017
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
A clonogenic common myeloid progenitor giving rise to all myeloid lineages was prospectively identified and purified in mouse by cell-surface markers and flow cytometry; it gives rise to either megakaryocyte/erythrocyte or granulocyte/macrophage progenitors, and commitment to these two fates was proposed to be mutually exclusive.
A clonogenic common myeloid progenitor that gives rise to all myeloid lineages · Akashi K, Traver D, Miyamoto T, Weissman IL · Nature 2000What this evidence cannot show
Mouse, and the mutual-exclusivity claim is a proposal in the paper's own framing rather than a demonstrated constraint. First-pass expression profiling of the sorted populations was by the methods of 2000. Read: abstract only.
Evidence that argues against this (1)
- Mutual exclusivity of GM and MegE commitment does not hold across human ontogeny. Fetal liver contains many oligopotent progenitors with intermingled myeloid, erythroid and megakaryocytic fates while adult marrow contains almost none, so the same fork is not the same fork at different ages. Since KMT2A-rearranged infant leukaemia arises prenatally, the mouse adult branch structure is the wrong reference for it. Read: abstract only. Notta F et al. 2016
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Integrating flow cytometric, transcriptomic and functional data at single-cell resolution shows that during homeostasis individual human HSCs gradually acquire lineage biases along multiple directions without passing through discrete hierarchically organised progenitor populations, with unilineage-restricted cells emerging directly from a continuum of low-primed undifferentiated haematopoietic stem and progenitor cells.
Human haematopoietic stem cell lineage commitment is a continuous process · Velten L, Haas SF, Raffel S, Blaszkiewicz S, Islam S, Hennig BP, et al., Steinmetz LM · Nature Cell Biology 2017What this evidence cannot show
Steady-state adult haematopoiesis; a continuum under homeostasis does not exclude discrete intermediates under stress, regeneration or in fetal life. Single-cell transcriptome depth in 2017 was limited and functional validation was on a subset. Read: abstract only.
Evidence that argues against this (1)
- The marker-defined progenitor populations the continuum model displaces remain reproducibly isolable and functionally distinct in vitro, and every clinical and experimental haematology protocol in use is built on them, including the founding MLL-AF9 experiment in which sorted committed GMP were transformed and gave a leukaemia stem cell with a GMP expression profile. A continuum at the population level is compatible with a sorted gate behaving as a coherent experimental unit. Read: abstract only. Krivtsov AV et al. 2006
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The human blood hierarchy is reshaped across ontogeny: fetal liver contains many oligopotent progenitors with intermingled myeloid, erythroid and megakaryocytic fates, whereas adult bone marrow contains almost none, so the same surface-marker gates do not isolate the same populations at different developmental stages.
Distinct routes of lineage development reshape the human blood hierarchy across ontogeny · Notta F, Zandi S, Takayama N, et al., Dror Y, Dick JE · Science 2016What this evidence cannot show
Single-cell and index-sorting functional assays in human fetal liver, cord blood and adult marrow; assay readouts constrain which fates can be detected, and absence of a fate in culture is weaker evidence than its presence. Read: abstract only.
Evidence that argues against this (1)
- If the hierarchy is ontogeny-dependent then arguments that KMT2A fusions avoid the MegE branch, which rest on adult-derived cell lines and adult-defined gates, may be describing the adult hierarchy rather than the fetal one in which the disease actually begins. The asymmetry recorded in this entry could be an artefact of where the observations were made. Read: abstract only (same paper). Notta F et al. 2016
GMP is one of the two branches downstream of the CMP gate
MEP is the other branch downstream of the CMP gate
- ← part ofMultipotent progenitor (MPP) — the stage where the undifferentiated KMT2A-r infant leukaemias map
CMP is the myeloid-restricted stage downstream of the MPP
zero basophil and mast-cell records sharpen the branch asymmetry recorded there
myelo-erythroid progenitor: required, proliferation promoted
- ← expressed inCD34 - named in 132 entries of this graph and written in none, and it does not mean what the graph has been using it to mean
MLL-rearranged AML arises from HSC and from CMP origins with contrasting chemoresistance, yet LSCs from BOTH carry a mature CD34-negative-to-low CD38-positive phenotype - functional origin and CD34 status come apart
REQUIRED IN: SMARCD2 controls EARLY steps in the differentiation of myeloid-erythroid progenitor cells in mice and zebrafish, with human patients showing neutropenia and myelodysplasia (PMID 28369036)