KITLG (stem cell factor) - the niche ligand of KIT, zero mentions anywhere in this graph, and its own symbol collides with the SCF ubiquitin ligase this corpus talks about constantly
also called SCF (stem cell factor) · KIT ligand · MGF · mast cell growth factor · steel factor
KITLG is the growth signal that marrow blood vessels and support cells hand to blood stem cells through the receptor KIT. In mice with MLL-AF9 leukaemia it keeps one of the two leukaemia stem cell types alive, and this graph had never mentioned it once.
The same thing, for a scientist
KITLG (UniProt P21583, 273 aa; the record's own entry name is SCF_HUMAN) is the ligand for KIT and, with CXCL12, one of the two canonical HSC-maintenance factors of the marrow niche [E1]. Ding and colleagues deleted *Scf* from each candidate source and found HSCs depleted only when it was removed from endothelial cells or from LEPR-expressing perivascular stromal cells - and NOT when removed from haematopoietic cells, osteoblasts, or nestin-cre/nestin-creER-expressing cells [E2]. In murine MLL-AF9 leukaemia, SCF (or FLT3L) promotes the survival of LMPP-like leukaemia stem cells through STAT5-mediated MCL1, while IL-3/IL-6 support GMP-like LSCs through STAT3/NF-kB-mediated BCL2, and AML cells cultured in SCF alone are more clonogenic and more leukemogenic [E3]. That result splits the LSC compartment by cytokine, and this graph holds `cell-type-lmpp` as a canonical id. KITLG is not a cell-intrinsic DepMap dependency [E4], as expected for a ligand supplied in trans. Its receptor already has an entry here (`gene-kit`); the ligand did not.
No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.
The symbol is a trap, and this corpus is the reason
UniProt's entry name for P21583 is SCF_HUMAN and the record carries SCF as a gene-name synonym. In this corpus SCF means something else entirely: a word-boundary grep over data/graph/nodes returns 41 files and 153 occurrences of SCF, and the contexts are the SKP1-CUL1-F-box ubiquitin ligase - 'MLL is degraded by SCF(Skp2) at S phase', 'ATF4 is an SCF-betaTrCP substrate', 'the fusion escapes SCF(Skp2)'. Two of the 153 are the SCFR string (KIT's own alias). So a corpus search for stem cell factor under the token SCF returns a literature about ubiquitin-mediated degradation of KMT2A, which is a central topic here and would read as a rich result. The same collision runs the other way in PubMed: (KITLG[tiab] OR "stem cell factor"[tiab] OR SCF[tiab]) intersected with the KMT2A/MLL half returns 5 records, of which one is a Cks1/Cks2 paper and one is 'The stem cell factor SALL4...' - where 'stem cell factor' is an English phrase describing a transcription factor, not this protein. Search this gene as KITLG or 'KIT ligand'; never as a bare SCF.
Which cell supplies it, and one result that sits awkwardly with an existing entry here
Ding and colleagues used Scf-gfp knock-in mice and conditional deletion from each candidate source. Scf was primarily expressed by perivascular cells throughout marrow; HSC frequency and function were unaffected by deletion from haematopoietic cells, osteoblasts, or nestin-cre or nestin-creER-expressing cells; HSCs were depleted by deletion from endothelial cells or from Lepr-expressing perivascular stromal cells, and most were lost when both were deleted 2.
This is a tension with cell-type-nestin-msc, and it is reported rather than acted on. That entry (Q-marrow-niche, cycle 40) records nestin+ BMSCs as supporting MLL-AF9 survival and chemotherapy relapse through OXPHOS and glutathione. E2 does not contradict that: it says nestin-cre-marked cells are not the required SCF source for normal HSCs, which is a different claim about a different factor in a different setting, and the Nes-GFP versus nestin-cre labelling discrepancy is a long-running dispute in the niche field. The two entries should be read together; no correction to cell-type-nestin-msc is requested.
| gene symbol | KITLG |
|---|---|
| uniprot | P21583 |
| length AA | 273 |
| accession resolution | gene_exact:KITLG AND organism_id:9606 AND reviewed:true returned exactly ONE reviewed human record, P21583, primary gene name read back as KITLG, 273 aa, synonyms MGF and SCF. Fetched live 2026-09-22. The record's ENTRY NAME is SCF_HUMAN, which is the origin of the collision recorded in collision_note - an agent reading the entry name rather than the primary gene name would file this protein under a symbol that means a ubiquitin ligase in this corpus. |
| localization | THREE COMPARTMENTS FOR THREE MOLECULAR SPECIES, and they are not interchangeable. UniProt P21583: isoform 1 is CELL MEMBRANE; isoform 2 is CYTOPLASM, cytoskeleton, cell membrane, and cell projections (lamellipodium, filopodium), the cytoplasm and projection terms carrying ECO:0000269; the processed 'Soluble KIT ligand' chain (26-190) is SECRETED. Membrane-anchored SCF signals persistently at a cell-cell contact while soluble SCF signals transiently and diffusibly - a distinction with real consequences for a niche model, and one that no KMT2A-relevant paper found in this beat addresses. |
| subnuclear localization | NOT APPLICABLE BY TOPOLOGY, argued from architecture rather than from a literature search. Isoform 1 and the processed chain are extracellular or membrane-anchored; isoform 2's non-membrane pool is annotated cytoplasmic and cytoskeletal, with no nuclear term and no chromatin-binding module in a 273-residue protein. This field does NOT claim nuclear KITLG was sought and excluded - no such search was run here, and UniProt's silence is not evidence (see the counter-evidence on E1). |
| subnuclear absence class | not-applicable-by-topology |
| cell type | SOURCE: endothelial cells and LEPR-expressing perivascular stromal cells, established by conditional deletion as the two compartments whose Scf is required for HSC maintenance, with osteoblasts, haematopoietic cells and nestin-cre-marked cells each dispensable [E2]. TARGET: HSCs via KIT, and in MLL-AF9 leukaemia specifically the LMPP-LIKE leukaemia stem cell, which SCF sustains while IL-3/IL-6 sustain the GMP-like LSC [E3]. This graph holds cell-type-lmpp, cell-type-hsc-bone-marrow and cell-type-sinusoidal-endothelial-cell, so all three compartments in that sentence already have entries. |
| cell state | AN LMPP-LIKE, HIGHLY CLONOGENIC, MCL1-DEPENDENT LEUKAEMIA STEM CELL STATE. In murine MLL-AF9, SCF signalling drives Stat5-mediated Mcl1 and AML cells grown in SCF alone are more clonogenic and more leukemogenic than the IL-3/IL-6-grown, Stat3/NF-kB/Bcl2-dependent GMP-like state [E3]. The distinction is drug-relevant rather than descriptive: this graph holds drug-venetoclax, a BCL2 inhibitor, and the SCF-supported compartment is the one defended by MCL1 instead. |
| epigenetic readout | NOT MEASURED, and the searchable literature is contaminated by the symbol. No histone-mark, DNA-methylation or menin-occupancy measurement at the KITLG locus in a KMT2A-rearranged system was found. The bound is weak and is stated as weak, because the obvious query token is unusable: a bare SCF search returns the SKP1-CUL1-F-box literature, and 'stem cell factor' as a phrase returns papers using it as ordinary English (E3's neighbour, PMID 28974232, is titled 'The stem cell factor SALL4...'). ABSENCE CLASS: measured-not-searched - the signal would be inside existing KMT2A-r ChIP-seq and RNA-seq depositions, and the reason it has not been found is that the search space excludes it on the symbol. |
| shown for fusion | YES for murine MLL-AF9, as a survival factor for the LMPP-like leukaemia stem cell compartment acting through STAT5-MCL1 [E3]. NO for any KMT2A-rearranged B-ALL experiment, NO for KMT2A::AFF1, NO for any human primary sample, and NO for any in vivo demonstration that the marrow niche is the SCF source in this disease - E3 supplies the cytokine in culture and E2's source mapping is in normal mice. |
| intact union membership | ABSENT from the 277-partner IntAct REST union of the six core baits (analyses/interactome-union-migrated-to-rest/results.json, generated 2026-09-22), checked on the symbol and on P21583. |
| collision note | NEW COLLISION, MEASURED HERE, NOT PREVIOUSLY IN THE REGISTRY. data/symbol-collisions.json was read first (2026-09-22, case-sensitive over d['collisions'], 177 rows): no row for SCF, KITLG, 'stem cell factor' or P21583. THE COLLISION: UniProt's entry name for KITLG is SCF_HUMAN and SCF is a curated gene-name synonym, while in THIS corpus a word-boundary grep for SCF returns 41 files and 153 occurrences, essentially all of them the SKP1-CUL1-F-box ubiquitin ligase - SCF(Skp2) degrading MLL, SCF-betaTrCP substrates - a central topic of this graph. Two of the 153 are the string SCFR, which is KIT's alias. In PubMed the same token misfires the other way: (KITLG[tiab] OR "stem cell factor"[tiab] OR SCF[tiab]) AND the KMT2A/MLL half returns 5 records including a Cks1/Cks2 ubiquitin paper and PMID 28974232, whose title uses 'stem cell factor' as English to describe SALL4. RULE: query this gene as KITLG or 'KIT ligand'. Corpus mentions of the token KITLG before this entry: ZERO. |
| corpus presence before this entry | KITLG was mentioned in 0 of the node files in this graph. The receptor gene-kit already had an entry; the ligand had no entry and no mention. |
domains — the numbers behind the chart
[
{
"name": "Signal peptide",
"start": 1,
"end": 25
},
{
"name": "Kit ligand (mature chain)",
"start": 26,
"end": 273
},
{
"name": "Soluble KIT ligand (processed chain)",
"start": 26,
"end": 190
},
{
"name": "Extracellular topological domain",
"start": 26,
"end": 214
},
{
"name": "Transmembrane (helical)",
"start": 215,
"end": 237
},
{
"name": "Cytoplasmic tail",
"start": 238,
"end": 273
}
]depmap essentiality
{
"depmap_common_essential": false,
"tissue_blocks": 29,
"n_screens": 1258,
"mean_gene_effect": 0.1029,
"min_gene_effect": -0.3885,
"max_gene_effect": 0.6734,
"n_lines_at_or_below_minus_1": 0,
"open_targets_data_version": "year 26, month 06",
"open_targets_api_version": "26.6.3",
"fetched": "2026-09-22",
"control": "Live zero-screen control in the same call and code path: MEN1 returned 29 tissue blocks, 1,258 screens, mean gene effect -0.3616, minimum -2.5262, 79 lines at or below -1.0 - so a zero count of dependent lines here is a measured zero, not an empty payload with a passing approvedSymbol readback. DepMap release recorded as Open Targets dataVersion year 26 month 06 (rule 83), API 26.6.3, fetched 2026-09-22. No median-split proportion is reported (rule 84). AND THE NULL IS AN ABSENCE BY CONSTRUCTION: DepMap screens cells in monoculture with no stroma, so a niche factor acting in trans cannot score however essential it is in vivo. The figure is recorded so nobody re-derives it and mistakes it for a finding."
}- SCF-KIT is the axis that maintains normal HSCs, mast cells, melanocytes and germ cells. Any attempt to remove SCF support from an LMPP-like LSC removes it from the patient's own stem cell pool at the same time - E2 is simultaneously the mechanism and the toxicity argument.
- Removing one cytokine circuit selects for the other. E3 shows two LSC populations sustained by different cytokines and different anti-apoptotic proteins; blocking SCF-STAT5-MCL1 alone predicts outgrowth of the IL-3/IL-6-BCL2 population, and this graph's venetoclax entry describes the mirror-image failure.
- The target is a ligand supplied by endothelium. Delivering an inhibitor to the endothelial and LEPR+ perivascular sources without collapsing marrow vascular function is an unsolved delivery problem, and no such agent exists.
- Is the LMPP-like/GMP-like LSC split of E3 present in KMT2A::AFF1 B-lineage disease, where the cell of origin is thought to be an LMPP-like or pro-B progenitor rather than a GMP? If it is, the SCF-STAT5-MCL1 circuit is the one that matters in the index disease and venetoclax targets the other arm.
- Is membrane-bound or soluble SCF the relevant species? No KMT2A paper found here distinguishes them, and they predict different spatial requirements for the niche.
- Does the KITLG locus carry any fusion-dependent chromatin signal? The question has never been asked, partly because the symbol is unsearchable in the obvious form.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The reviewed human KITLG record is UniProt P21583 (entry name SCF_HUMAN), 273 aa, with signal peptide 1-25, a membrane-anchored form (extracellular 26-214, transmembrane 215-237, cytoplasmic 238-273) and a soluble processed chain 'Soluble KIT ligand' at 26-190; it is the ligand for the receptor tyrosine kinase KIT and forms a heterotetramer binding two KIT molecules.
UniProtKB P21583 (SCF_HUMAN) - Kit ligand · UniProtKB 2026What this evidence cannot show
Resolved by gene_exact:KITLG AND organism_id:9606 AND reviewed:true, exactly ONE reviewed human record; primary gene name read back as KITLG, 273 aa, synonyms MGF and SCF. Fetched 2026-09-22. Note the isoform structure matters and is not decoration: isoform 1 is cell-membrane, isoform 2 is cytoplasmic/lamellipodial/filopodial, and the soluble chain is secreted - membrane-bound and soluble SCF are functionally different ligands, and no KMT2A paper distinguishes them.
Evidence that argues against this (1)
- A curated feature table records what curators accepted, not what was sought and excluded, so no absence claim follows from it. The worked example is in this beat: UniProt P61073 annotates no nuclear localisation signal for CXCR4 while a mapped, mutated, functionally tested NLS at 146-RPRK-149 is published. Bao Y et al. 2019
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In mouse, Scf is primarily expressed by perivascular cells throughout bone marrow; HSC frequency and function were unaffected by conditional Scf deletion from haematopoietic cells, osteoblasts, or nestin-cre or nestin-creER-expressing cells, but HSCs were depleted by deletion from endothelial cells or from Lepr-expressing perivascular stromal cells, with most lost when both were deleted.
Endothelial and perivascular cells maintain haematopoietic stem cells · Ding L, Saunders TL, Enikolopov G, Morrison SJ · Nature 2012What this evidence cannot show
Mouse, normal haematopoiesis, no leukaemia. Conditional-deletion results are only as specific as the Cre driver, and the nestin-cre arm in particular sits inside an unresolved field-wide dispute about how faithfully nestin-cre and Nes-GFP mark the same perivascular population - so the negative arm of this experiment is weaker than the positive arms. Read from the abstract.
Evidence that argues against this (1)
- This graph's cell-type-nestin-msc records nestin+ BMSCs as functionally supporting MLL-AF9 leukaemia through oxidative phosphorylation and glutathione-mediated antioxidant defence, and nestin+ cell depletion delaying leukaemogenesis. A reader taking E2's nestin-cre negative at face value would wrongly conclude the nestin+ compartment is inert; the two results concern different factors, different drivers and different diseases. Xiao P et al. 2018
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In murine MLL-AF9 AML, SCF or FLT3L promotes the survival of LMPP-like leukaemia stem cells by stimulating Stat5-mediated Mcl1 expression, whereas IL-3 or IL-6 supports GMP-like LSCs through Stat3/NF-kB-mediated Bcl2; AML cells cultured in SCF or Flt3L alone are more clonogenic in vitro and more leukemogenic in vivo than cells cultured in IL-3 and IL-6.
Distinct roles of hematopoietic cytokines in the regulation of leukemia stem cells in murine MLL-AF9 leukemia · Li Y, Seet CS, Mack R, et al. · Stem Cell Reports 2024What this evidence cannot show
Murine myeloid MLL-AF9, not infant B-lineage KMT2A::AFF1 (rule 73). The SCF here is supplied as a culture cytokine, so the experiment establishes that the LMPP-like LSC compartment DEPENDS on SCF signalling, not that the marrow niche supplies it in vivo in this disease - the in vivo source question is untouched. Read from the abstract.
Evidence that argues against this (1)
- The same paper's own framing cuts against a single-cytokine strategy: two LSC populations coexist and each is sustained by a different cytokine-to-anti-apoptotic-protein circuit, so removing SCF signalling selects for the IL-3/IL-6-supported GMP-like LSC rather than eliminating the compartment. The authors argue for co-inhibition of both circuits for that reason. Li Y et al. 2024
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
KITLG is not a cell-intrinsic dependency in DepMap: 29 tissue blocks, 1,258 screens, mean gene effect +0.1029, range -0.3885 to 0.6734, zero lines at or below a gene effect of -1.0.
Open Targets Platform GraphQL - target.depMapEssentiality for ENSG00000049130 (KITLG), with meta.dataVersion · Open Targets Platform / DepMap 2026What this evidence cannot show
Live zero-screen control in the same call and code path: MEN1 returned 29 tissue blocks, 1,258 screens, mean gene effect -0.3616, minimum -2.5262, 79 lines at or below -1.0 - so a zero count of dependent lines here is a measured zero, not an empty payload with a passing approvedSymbol readback. DepMap release recorded as Open Targets dataVersion year 26 month 06 (rule 83), API 26.6.3, fetched 2026-09-22. No median-split proportion is reported (rule 84). AND THE NULL IS AN ABSENCE BY CONSTRUCTION: DepMap screens cells in monoculture with no stroma, so a niche factor acting in trans cannot score however essential it is in vivo. The figure is recorded so nobody re-derives it and mistakes it for a finding.
Evidence that argues against this (1)
- The same monoculture assay returns an equally flat null for CXCR4 (mean +0.0399, zero lines at or below -1.0) while an in vivo syngeneic MLL-AF9 CRISPR screen made CXCR4 a TOP cell-surface dependency whose deletion eradicated leukaemia. That pair is a direct demonstration that a DepMap null for a niche gene is uninformative about in vivo requirement. Ramakrishnan R et al. 2020
KITLG is the ligand of KIT; heterotetramer binding two KIT molecules
one of the two canonical HSC-maintenance factors of the niche, with CXCL12
endothelial Scf deletion depletes HSCs
SCF sustains LMPP-like leukaemia stem cells via STAT5-MCL1 in murine MLL-AF9
Scf is primarily expressed by perivascular cells throughout marrow
Scf deletion from Lepr-expressing perivascular cells depletes HSCs - LEPR marks the KITLG source
measures the SCF symbol collision recorded in that entry