Niche biology HAS been measured in KMT2A-rearranged disease - but in the index fusion itself there is essentially one marrow-niche molecular experiment, from 2011
also called niche specificity census · is the niche literature borrowed
We counted, molecule by molecule, how much of the bone-marrow niche literature has actually been done in this particular leukaemia rather than borrowed from leukaemia in general. There is real work in KMT2A-rearranged disease as a class, but for the exact infant fusion this project studies, almost none of it exists.
The same thing, for a scientist
A per-molecule PubMed census run 2026-09-22 with querytranslation read back on every query, one tag per term (rule 94), and rule-67 controls through the identical shape. **The answer is not a simple yes or no and the beat brief's framing - that the niche literature is borrowed - is half right.** Against the disease half `KMT2A[tiab] OR MLL-rearranged[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab]` (2,922 records), a generic niche half returns **120** records, so niche work in KMT2A-rearranged disease exists at scale [E1]. Per molecule the counts are small but mostly non-zero: CXCR4 11, CXCL12 6, KITLG/SCF 5 (contaminated), SPP1/osteopontin 4, VCAM1 2, LEPR/leptin 2, ITGA4 1, SELE/E-selectin/uproleselan **0** [E1]. **But scope the disease half to the INDEX FUSION** - `MLL-AF4[tiab] OR KMT2A-AFF1[tiab] OR "t(4;11)"[tiab]`, 1,050 records - and its intersection with any niche term is **20**, of which the titles show only one reporting a marrow-niche molecular measurement: Orlovsky 2011, the RS4;11 MEIS1/HOXA knockdown that hampered the CXCR4/SDF-1 axis [E2, E3]. Two more are CNS-niche papers, a different anatomical compartment [E4]. So: measured in the CLASS, borrowed for the FUSION.
The census table, with the exact queries
Disease half throughout, unless stated: KMT2A[tiab] OR MLL-rearranged[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab], 2,922 records, translation read back as four separate [Title/Abstract] terms. All queries 2026-09-22.
| niche molecule | query term(s) | intersection | note |
|---|---|---|---|
| CXCR4 | CXCR4[tiab] | 11 | best-covered; includes the MLL-AF9 CRISPR screen and the infant MLL-R ALL stromal-protection paper |
| CXCL12 | CXCL12[tiab] | 6 | |
| KITLG | KITLG OR "stem cell factor" OR SCF | 5 | CONTAMINATED - includes a Cks1/Cks2 ubiquitin paper and one using 'stem cell factor' as English for SALL4 |
| SPP1 | SPP1 OR osteopontin | 4 | one is title-only and unreadable |
| VCAM1 | VCAM1[tiab] | 2 | VCAM-1[tiab] against the same half returns 0 - a hyphenation false zero |
| LEPR | LEPR OR leptin | 2 | |
| ITGA4 | ITGA4[tiab] | 1 | the 2026 natalizumab paper, and nothing else, ever |
| SELE | SELE OR E-selectin OR uproleselan | 0 | against an 8,608-record E-selectin literature |
| any niche term | niche OR stroma OR stromal OR microenvironment | 120 | |
| menin x CXCR4 | menin[tiab] AND CXCR4[tiab] | 0 | operands 1,264 and 17,210 |
Rule-67 controls, through the identical two-term AND shape: LEPR 2, SPP1 4, VCAM1 2, ITGA4 1, and menin[tiab] AND HOXA9[tiab] 33. The AND composes; the zeros for SELE and for menin x CXCR4 are real zeros and not non-composition artefacts.
The finding that matters: the class is not the fusion
MLL-AF4[tiab] OR KMT2A-AFF1[tiab] OR "t(4;11)"[tiab] returns 1,050 records - a substantial index-fusion literature. Intersected with any of niche, stroma, stromal, microenvironment, CXCR4, CXCL12, ITGA4 or VCAM1 it returns 20, and the titles were read rather than counted 3. What they are:
- One marrow-niche molecular measurement: Orlovsky 2011, RS4;11, MEIS1/HOXA7/9/10 knockdown hampering the CXCR4/SDF-1 axis 2.
- Two CNS-niche papers, a different anatomical compartment which this graph covers separately 4.
- Seventeen incidental matches: an siRNA nanoparticle delivery study, an immunoproteasome inhibitor study, ribavirin/EIF4E in infant ALL, a lineage-commitment paper, ex vivo drug-response profiling, adult LIC xenograft biology, a review of cellular origin, and similar. In these the niche words appear in passing.
So the honest answer to 'is niche retention or adhesion different in KMT2A-rearranged disease specifically' is three-layered:
- In KMT2A-rearranged disease as a class - YES, and it is measured, not assumed. Primary infant MLL-rearranged ALL blasts are protected by stroma differentially relative to non-MLL-rearranged ALL, and CXCR4 antagonism reverses it 5. That is a direct comparative claim in the right lineage and it is the single most important citation in this beat.
- In the index fusion KMT2A::AFF1 - essentially NO. One 2011 knockdown paper. Everything else in this graph's niche layer is MLL-AF9 mouse myeloid or MLL-rearranged-as-a-class.
- For individual molecules - mostly NO, and for E-selectin absolutely not. ITGA4 has exactly one paper in fourteen years of the symbol's existence, and it is from 2026.
What the census does NOT establish
Every count here is bounded by titles and abstracts (rule 80), so a niche measurement inside a supplementary figure is invisible - and this project has a worked case where exactly that happened. The counts are also bounded by the name lists, which are written out in the table so a reader can judge them; CD62E, GMI-1271, VLA-4, CD49d, SDF-1 and OB-R were not used as query terms, and adding them would raise some counts. And the census measures reporting, not measurement: RNA-seq and single-cell data from KMT2A::AFF1 models exist in GEO in which these transcripts were measured and never analysed, which is why the companion absence entry is classed measured-not-analysed rather than never-measured.
| script | ad hoc NCBI eutils esearch, db=pubmed, querytranslation read back on every query; full parameters in the entry body |
|---|---|
| status | run |
| run date | 2026-09-22 |
| question | Is niche retention or adhesion measured in KMT2A-rearranged disease specifically, or is this a general leukaemia literature being borrowed? |
| answer | BOTH, at different scopes. In KMT2A-rearranged disease as a CLASS the niche has been measured and the key comparative result is positive - primary infant MLL-rearranged ALL is differentially protected by stroma relative to non-MLL-rearranged ALL and CXCR4 antagonism reverses it. In the INDEX FUSION KMT2A::AFF1 there is essentially one marrow-niche molecular experiment, from 2011. Per molecule the counts are 11 (CXCR4) down to 0 (SELE). |
| rule 67 controls | Run through the IDENTICAL two-term AND shape against the same disease half: LEPR/leptin 2, SPP1/osteopontin 4, VCAM1 2, ITGA4 1; and menin[tiab] AND HOXA9[tiab] 33 for the menin queries. The AND composes, so the SELE zero and the menin-CXCR4 zero are measured zeros. |
| instrument failures observed | FOUR queries in this session returned an EMPTY querytranslation with no count field - CXCR4 AND the disease half, CXCR4 AND infant AND leukemia, SPP1/osteopontin AND the disease half, and SELE AND the disease half. All four returned normal non-zero or zero results with valid translations when re-run within a minute. An empty querytranslation is an instrument failure and NOT a zero; had any of the four been recorded as written, this census would have reported false zeros for CXCR4 and for SPP1, the two best-covered molecules in it. |
| name lists not searched | CD62E, GMI-1271, VLA-4, CD49d, SDF-1, OB-R, MGF, PBSF, CD106, CD184. Adding these would raise several counts and the bound is stated as a bound. |
| bounded rejections | The 17 incidental index-fusion hits are recorded as considered and dismissed at title level, so a future cycle can tell them from papers nobody looked for. |
disease half
{
"query": "KMT2A[tiab] OR MLL-rearranged[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab]",
"querytranslation": "\"KMT2A\"[Title/Abstract] OR \"MLL-rearranged\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract]",
"count": 2922
}index fusion half
{
"query": "MLL-AF4[tiab] OR KMT2A-AFF1[tiab] OR \"t(4;11)\"[tiab]",
"querytranslation": "\"MLL-AF4\"[Title/Abstract] OR \"KMT2A-AFF1\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract]",
"count": 1050
}per molecule counts — the numbers behind the chart
[
{
"molecule": "CXCR4",
"terms": "CXCR4[tiab]",
"operand_alone": 17210,
"intersection_with_disease_half": 11
},
{
"molecule": "CXCL12",
"terms": "CXCL12[tiab]",
"operand_alone": 7568,
"intersection_with_disease_half": 6
},
{
"molecule": "KITLG",
"terms": "KITLG[tiab] OR \"stem cell factor\"[tiab] OR SCF[tiab]",
"intersection_with_disease_half": 5,
"note": "contaminated by the SKP1-CUL1-F-box sense of SCF and by 'stem cell factor' as English"
},
{
"molecule": "SPP1",
"terms": "SPP1[tiab] OR osteopontin[tiab]",
"intersection_with_disease_half": 4
},
{
"molecule": "VCAM1",
"terms": "VCAM1[tiab]",
"intersection_with_disease_half": 2,
"note": "VCAM-1[tiab] against the identical half returns 0 - a hyphenation false zero"
},
{
"molecule": "LEPR",
"terms": "LEPR[tiab] OR leptin[tiab]",
"intersection_with_disease_half": 2
},
{
"molecule": "ITGA4",
"terms": "ITGA4[tiab]",
"intersection_with_disease_half": 1,
"note": "the single record is PMID 42345361, published 2026"
},
{
"molecule": "SELE",
"terms": "SELE[tiab] OR E-selectin[tiab] OR uproleselan[tiab]",
"operand_alone": 8608,
"intersection_with_disease_half": 0
},
{
"molecule": "any niche term",
"terms": "niche[tiab] OR stroma[tiab] OR stromal[tiab] OR microenvironment[tiab]",
"intersection_with_disease_half": 120
},
{
"molecule": "menin x CXCR4",
"terms": "menin[tiab] AND CXCR4[tiab]",
"operand_alone": 1264,
"intersection_with_disease_half": 0
}
]index fusion intersection
{
"query": "(MLL-AF4[tiab] OR KMT2A-AFF1[tiab] OR \"KMT2A::AFF1\"[tiab] OR \"t(4;11)\"[tiab]) AND (niche[tiab] OR stroma[tiab] OR stromal[tiab] OR microenvironment[tiab] OR CXCR4[tiab] OR CXCL12[tiab] OR ITGA4[tiab] OR VCAM1[tiab])",
"count": 20,
"disposition": "titles read: 1 marrow-niche molecular measurement (PMID 21518888), 2 CNS-niche (PMID 41371538, PMID 42139054), 17 incidental"
}- A low count is not a licence to claim novelty. Several of these molecules have small counts because the question is hard or the reagents are lacking, not because nobody thought of it, and a proposal built on 'nobody has done this' should say which of the two it thinks is true.
- The strongest result in this census is a 2013 paper whose potentiated partner drug, lestaurtinib, was abandoned. Building a clinical argument on it requires re-demonstrating the protection with an agent that will actually be used.
- Would the 2013 differential-protection experiment reproduce with a defined KMT2A::AFF1 cohort and a current therapeutic partner? It is the load-bearing result for this whole layer and it is thirteen years old and unreplicated.
- How many of the 120 generic niche records in KMT2A-rearranged disease carry a niche measurement in a supplementary figure that no abstract-level census can see?
- Do E-selectin ligands exist on KMT2A::AFF1 B-precursor blasts? The only true zero in this census is also the target with the best therapeutic-index claim in the wider AML literature.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Against a disease half of KMT2A[tiab] OR MLL-rearranged[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab] (2,922 records), per-molecule intersections are: CXCR4 11, CXCL12 6, KITLG/stem cell factor/SCF 5, SPP1/osteopontin 4, VCAM1 2, LEPR/leptin 2, ITGA4 1, SELE/E-selectin/uproleselan 0, and a generic niche/stroma/stromal/microenvironment half 120.
NCBI E-utilities esearch, db=pubmed - niche-molecule census with querytranslation readback and rule-67 controls · NCBI PubMed 2026What this evidence cannot show
All counts 2026-09-22 with querytranslation read back and one tag per term (rule 94). Bounded over titles and abstracts (rule 80). The KITLG count of 5 is known-contaminated: 'SCF' matches the SKP1-CUL1-F-box ubiquitin ligase and 'stem cell factor' matches ordinary English usage, so the true KITLG-specific count is lower. Name lists are incomplete by design and are written out in the entry body so the bound is judgeable. Two queries in this session returned an EMPTY querytranslation with no count - a transient NCBI failure, not a zero - and were re-run; any count reported here comes from a response with a non-empty translation.
Evidence that argues against this (1)
- A per-molecule count is a poor proxy for whether the biology has been studied, because one strong paper outweighs ten weak ones: ITGA4 scores 1 and that single paper is a 2026 Blood study with patient-derived xenografts and an approved antibody, while CXCR4 scores 11 including a weightlessness-osteoporosis sequencing study and a Treg migration paper. The counts are a map of the literature's shape, not a ranking of evidence strength. Rubio-Gayarre A et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the index fusion literature (MLL-AF4[tiab] OR KMT2A-AFF1[tiab] OR "t(4;11)"[tiab], 1,050 records), the intersection with any of niche, stroma, stromal, microenvironment, CXCR4, CXCL12, ITGA4 or VCAM1 is 20 records, and on reading the titles only one reports a marrow-niche molecular measurement - the RS4;11 MEIS1/HOXA knockdown that hampered the CXCR4/SDF-1 axis.
Down-regulation of homeobox genes MEIS1 and HOXA in MLL-rearranged acute leukemia impairs engraftment and reduces proliferation · Orlovsky K, Kalinkovich A, Rozovskaia T, et al. · Proc Natl Acad Sci U S A 2011What this evidence cannot show
The 20 titles were read; the abstracts of the 17 judged incidental were not, so a niche measurement reported in an abstract whose title does not signal it could have been missed. Note also that the query translation normalised "KMT2A::AFF1" to "KMT2A-AFF1", making one of the four index-fusion terms a duplicate - which does not change the count but does mean the colon-form spelling was never independently tested. The cytogenetic term was QUOTED throughout (rule 95); an unquoted t(4;11) inside a group makes PubMed invent a UID filter and return a false zero.
Evidence that argues against this (1)
- Scoping to the index fusion may itself hide the answer, which is this project's rule 77. The strongest lineage-appropriate result - differential stromal protection of primary infant MLL-rearranged ALL reversed by CXCR4 antagonism - is written as 'MLL-R' rather than as t(4;11) and would be excluded by the index-fusion scoping used here, even though its samples are overwhelmingly likely to include t(4;11) cases. Sison EA et al. 2013
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
Of the 20 index-fusion niche-term intersections, the titles comprise one marrow-niche molecular measurement, two CNS-niche adaptation papers, and seventeen incidental matches including siRNA nanoparticle delivery, immunoproteasome inhibition, ribavirin/EIF4E targeting, lineage commitment, ex vivo drug-response profiling, adult LIC xenograft biology and a review of cellular origin.
NCBI E-utilities esearch, db=pubmed - niche-molecule census with querytranslation readback and rule-67 controls · NCBI PubMed 2026What this evidence cannot show
This is a title-level disposition by one reader in one session and is the weakest item in the entry. It is recorded as a BOUNDED REJECTION rather than left invisible: someone re-running this census will see which records were considered and dismissed, and can disagree. Titles read but abstracts not opened for the seventeen.
Evidence that argues against this (1)
- Title-level triage has a known failure mode in this project: a paper whose title looks squarely off-beat can carry the measurement in a supporting panel, and rule 80 was written after exactly that. Any of the seventeen could contain a niche measurement not signalled in its title. Orlovsky K et al. 2011
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Two of the twenty index-fusion niche-term records concern the CNS niche rather than the marrow niche, including a 2026 report that mixed-lineage leukaemia cells undergo unique adaptations in the CNS niche.
Mixed-lineage leukemia cells undergo unique adaptations in the CNS niche · Exp Hematol 2026What this evidence cannot show
Title-level identification; the abstract was not read in this beat. The CNS niche is a different anatomical compartment from the marrow niche and this graph covers it separately (cell-state-cns-niche-adapted, and the CNS deferral in ana-marrow-niche-rejected-candidates). It is counted out of the marrow census rather than against it.
Evidence that argues against this (1)
- Treating the CNS niche as out of scope is a choice, and the two compartments are mechanistically linked in this graph's own entries: VCAM1-VLA-4 blockade attenuated intracerebral invasion in MLL-AF9 mouse AML, so an adhesion molecule counted in the marrow census has its strongest KMT2A-relevant phenotype in the CNS. Oncogene 2022
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In primary infant MLL-rearranged ALL, stroma differentially protects blasts from spontaneous apoptosis relative to non-MLL-rearranged ALL, stroma-produced soluble factors alone promote proliferation and cell-cycle entry, stroma protects both bulk blasts and LSCs from lestaurtinib, and CXCR4 inhibition with AMD3100 markedly enhanced lestaurtinib efficacy in an MLL-R ALL xenograft.
MLL-rearranged acute lymphoblastic leukaemia stem cell interactions with bone marrow stroma promote survival and therapeutic resistance that can be overcome with CXCR4 antagonism · Sison EA, Rau RE, McIntyre E, Li L, Small D, Brown P · Br J Haematol 2013What this evidence cannot show
MLL-rearranged as a class in human primary infant samples; the abstract does not break the cohort down by fusion partner. This is the single best answer in the literature to the beat's central question and it is a 2013 paper that has not, as far as this census found, been repeated with modern tools, a defined fusion breakdown, or a current therapeutic partner. Read from the abstract.
Evidence that argues against this (1)
- The differential-protection claim is comparative and its comparator is 'non-MLL-rearranged ALL', a heterogeneous group. A difference against a mixed comparator does not identify what about KMT2A rearrangement produces it, and no subsequent study has decomposed it. Sison EA et al. 2013
census of how much niche biology is measured in this disease rather than borrowed
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
one marrow-niche molecular measurement in 1,050 index-fusion papers
the only true zero in the census
one record in the entire literature, published 2026
differential stromal protection of primary infant MLL-rearranged ALL is the load-bearing comparative result
re-measures and extends the cycle-40 rejection ledger's counts
the index disease whose niche literature this census bounds
the rejection half of the same census