Nobody has reported what a menin inhibitor does to CXCR4: zero papers name menin and CXCR4 together, and the RNA-seq that would answer it is already public
also called menin inhibitor CXCR4 absence · niche axis under menin inhibition
The drug class this whole project is about collapses the gene programme that makes leukaemia cells respond to the marrow's hold-still signal, and no one has ever checked whether the receptor for that signal goes down too. The data to check it is already sitting in a public archive.
The same thing, for a scientist
A bounded absence with a named closing route. `menin[tiab] AND CXCR4[tiab]` returns **0 records**, with a valid querytranslation, an explicit 'No items found.', and operands of 1,264 and 17,210; the rule-67 positive control through the identical shape, `menin[tiab] AND HOXA9[tiab]`, returns 33 [E1]. Widening to the drug-name list per rule 72 - revumenib, ziftomenib, VTP-50469, VTP50469, MI-503, SNDX-5613, plus 'menin inhibitor', 'menin inhibitors' and 'menin inhibition' - and intersecting with CXCR4, CXCL12, SDF-1, niche, stroma or stromal returns **3 records, none on topic**: a GHRH review, the ziftomenib first-approval review, and a gastrointestinal **stromal** tumour paper that matched on the disease name [E2]. The prediction being untested is sharp: menin inhibition down-regulates HOXA9/MEIS1 in KMT2A-rearranged AML [E3], MEIS1 drives CXCR4 function through SYTL1-controlled receptor trafficking [E4], and in RS4;11 (KMT2A::AFF1) knockdown of MEIS1 or HOXA7/9/10 hampers the CXCR4/SDF-1 axis [E5]. ABSENCE CLASS: **measured-not-analysed** - GEO holds deposited human menin-inhibitor transcriptomes in KMT2A-rearranged AML, including GSE281506 (42 samples, expression profiling by high-throughput sequencing) [E6], so CXCR4's response is in a public matrix and has simply never been read out.
Why the class is measured-not-analysed and not never-measured
The distinction is the difference between years of wet work and an afternoon of compute, which is why the standing rule requires an archive query before an absence is bounded by literature alone. One db=gds query settles it: 88 GEO records match 'revumenib' and 259 match 'menin inhibitor', and inspecting the top hits gives GSE281506, 'Combining Menin and MEK inhibition to target poor prognosis KMT2A-rearranged RAS pathway-mutant acute myeloid leukemia', Homo sapiens, 42 samples, expression profiling by high throughput sequencing 6. The paper that establishes the HOXA9/MEIS1 collapse under menin inhibition ran RNA-seq of its own and reported HDAC9 out of it 3 - the CXCR4 row was in that matrix too and nobody looked at it.
So the correct statement is not 'nobody has measured this'. It is nobody has reported it, and the measurement already exists in public data.
What the queries were, exactly
All run 2026-09-22 against NCBI esearch with the querytranslation read back, one tag per term (rule 94):
| query | translation | count |
|---|---|---|
menin[tiab] | "menin"[Title/Abstract] | 1,264 |
CXCR4[tiab] | "CXCR4"[Title/Abstract] | 17,210 |
menin[tiab] AND CXCR4[tiab] | "menin"[Title/Abstract] AND "CXCR4"[Title/Abstract] | 0 |
menin[tiab] AND HOXA9[tiab] (rule 67 control, identical shape) | "menin"[Title/Abstract] AND "HOXA9"[Title/Abstract] | 33 |
| drug-name list (9 spellings) | as written out in E2 | 344 |
| drug-name list AND (CXCR4 OR CXCL12 OR SDF-1 OR niche OR stroma OR stromal) | as written out in E2 | 3, all off-topic |
A retrieval trap worth recording: the token stroma/stromal in a niche query matches gastrointestinal STROMAL tumour, and one of the three hits above is exactly that. A beat that counted 3 and did not read the titles would have recorded the absence as partially closed.
| absence class | measured-not-analysed |
|---|---|
| what is absent | Any reported measurement of CXCR4 (or CXCL12, or SDF-1-directed migration) in a KMT2A-rearranged system under pharmacological menin inhibition. |
| what would close it | Re-analysis of deposited menin-inhibitor transcriptomes in KMT2A-rearranged cells for CXCR4 and SYTL1, plus a surface-CXCR4 flow readout and an SDF-1 transwell in a KMT2A::AFF1 line under revumenib. Compute plus one bench week, no new cohort. |
| script | ana-cxcr4-under-menin-inhibition-geo-reanalysis |
| status | run |
| run date | 2026-09-22 |
| archive check | GEO db=gds, 2026-09-22: 'revumenib' 88 records, 'menin inhibitor' 259 records, 'VTP-50469' 701 (loose match, treated as unreliable). GSE281506 confirmed by esummary: expression profiling by high throughput sequencing, 42 samples, Homo sapiens, 'Combining Menin and MEK inhibition to target poor prognosis KMT2A-rearranged RAS pathway-mutant acute myeloid leukemia'. This is what moves the class from never-measured to measured-not-analysed. |
| name list not searched | bleximenib, enzomenib, MI-2-2, MI-538, DS-1594, JNJ-75276617, BMF-219/icovamenib. Naming them is the honest bound: adding them could close this absence and the search should be re-run with them before anyone builds on the zero. |
| retrieval trap found | The token stroma/stromal in a niche query matches GASTROINTESTINAL STROMAL TUMOR. One of the three hits in the widened query is exactly that, so a beat counting 3 without reading titles would have recorded this absence as partially closed by a GIST paper. |
queries — the numbers behind the chart
[
{
"query": "menin[tiab]",
"querytranslation": "\"menin\"[Title/Abstract]",
"count": 1264
},
{
"query": "CXCR4[tiab]",
"querytranslation": "\"CXCR4\"[Title/Abstract]",
"count": 17210
},
{
"query": "menin[tiab] AND CXCR4[tiab]",
"querytranslation": "\"menin\"[Title/Abstract] AND \"CXCR4\"[Title/Abstract]",
"count": 0,
"warninglist": "outputmessages: No items found.; phrasesignored: []; quotedphrasesnotfound: []"
},
{
"query": "menin[tiab] AND HOXA9[tiab] (RULE 67 POSITIVE CONTROL, identical shape)",
"querytranslation": "\"menin\"[Title/Abstract] AND \"HOXA9\"[Title/Abstract]",
"count": 33
},
{
"query": "revumenib[tiab] OR ziftomenib[tiab] OR VTP-50469[tiab] OR VTP50469[tiab] OR MI-503[tiab] OR SNDX-5613[tiab] OR \"menin inhibitor\"[tiab] OR \"menin inhibitors\"[tiab] OR \"menin inhibition\"[tiab]",
"count": 344
},
{
"query": "(the nine-name list) AND (CXCR4[tiab] OR CXCL12[tiab] OR SDF-1[tiab] OR niche[tiab] OR stroma[tiab] OR stromal[tiab])",
"count": 3,
"disposition": "all three read and all three off-topic: PMID 42346642 (GHRH antagonist review), PMID 41653248 (Ziftomenib: First Approval), PMID 35499757 (MOZ and Menin-MLL complexes in GASTROINTESTINAL STROMAL TUMOR - matched on the word stromal in the disease name)"
}
]- If menin inhibition does lower CXCR4 function, then the drug is partly a mobilising agent, and combining it with a CXCR4 antagonist is redundant at best and additively myelosuppressive at worst.
- A negative re-analysis would not be a clean negative: E4's caveat means CXCR4 function can fall through SYTL1-dependent trafficking with the transcript unchanged, so a transcriptome-only result must not be reported as refuting the prediction.
- Does CXCR4 transcript move under revumenib in a KMT2A::AFF1 line, and if not, does SYTL1?
- Would a menin inhibitor be doing part of its work by de-adhering the blast, and would that be visible as an early rise in circulating blasts before marrow clearance?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No paper names menin and CXCR4 together in its title or abstract: menin[tiab] AND CXCR4[tiab] returns 0 records with a valid querytranslation and an explicit 'No items found.', against operands of 1,264 and 17,210, while the identical-shape positive control menin[tiab] AND HOXA9[tiab] returns 33.
NCBI E-utilities esearch, db=pubmed - bounded-absence queries with querytranslation readback and a rule-67 positive control · NCBI PubMed 2026What this evidence cannot show
Bounded over TITLES AND ABSTRACTS only (rule 80). A CXCR4 row inside a menin-inhibitor RNA-seq supplementary table, or a CXCR4 track in a deposited CUT&RUN dataset, is invisible to this query - which is the whole reason this entry's absence class is measured-not-analysed rather than never-measured. Queried 2026-09-22.
Evidence that argues against this (1)
- A zero in an index that covers only part of a document cannot bound a claim about the whole document, and this graph has a worked case with a named paper: a bounded SET-NUP214 query returned 0 for H3K79 with clean positive controls and four spellings, while an H3K79me2 ChIP across five HOXA promoters sits inside this paper as a supporting figure and is invisible to any abstract-level search. The same escape hatch is open here and is named rather than closed. Van Vlierberghe P et al. 2008
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Widening to nine menin-inhibitor name spellings (revumenib, ziftomenib, VTP-50469, VTP50469, MI-503, SNDX-5613, 'menin inhibitor', 'menin inhibitors', 'menin inhibition') gives 344 records; intersecting that list with CXCR4, CXCL12, SDF-1, niche, stroma or stromal gives 3 records, and all three are off-topic - a growth-hormone-releasing-hormone review, the ziftomenib first-approval review, and a gastrointestinal stromal tumour study that matched on the word 'stromal' in its disease name.
NCBI E-utilities esearch, db=pubmed - bounded-absence queries with querytranslation readback and a rule-67 positive control · NCBI PubMed 2026What this evidence cannot show
Rule 72 compliance: singular, plural and nominalisation were all included, and hyphenated and unhyphenated compound codes were treated as different strings. The name list is written out in full here so a reader can judge the bound - bleximenib, enzomenib, MI-2-2, MI-538, DS-1594 and JNJ-75276617 were NOT included and are the obvious way this bound could be too tight. The three hits were identified by fetching their titles, not by counting.
Evidence that argues against this (1)
- The bare-operand check rule 72 demands cuts the other way here and strengthens rather than weakens the absence: menin[tiab] alone returns 1,264 and its intersection with CXCR4 is still 0, so the zero is not an artefact of the drug-name list. NCBI PubMed 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In KMT2A-rearranged AML, menin-inhibitor monotherapy down-regulates HOXA9/MEIS1, and RNA-sequencing of venetoclax plus menin inhibitor identified HDAC9 as specifically repressed by the combination; hypoxia induction sensitised KMT2A-rearranged AML to venetoclax plus MI-503.
Repressing HIF-1alpha-induced HDAC9 contributes to the synergistic effect of venetoclax and MENIN inhibitor in KMT2Ar AML · Ling Q, Zhou Y, Qian Y, et al. · Biomark Res 2023What this evidence cannot show
KMT2A-rearranged AML cell lines, primary cells ex vivo and a MOLM13 xenograft - myeloid, not infant B-lineage KMT2A::AFF1 (rule 73). This entry cites it for two narrow facts read directly from the fetched abstract: that menin inhibition collapses HOXA9/MEIS1 in this disease, and that transcriptome-wide RNA-seq under menin inhibition exists. It does not cite it for anything about CXCR4, which the paper does not mention.
Evidence that argues against this (1)
- The same abstract records that the BCL-2/BCL-XL and HOXA9/MEIS1 mechanisms were 'not further strengthened by their combination' - a caution that a transcriptional output collapsing under monotherapy does not license predicting a combination effect on any particular downstream gene, including CXCR4. Ling Q et al. 2023
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MEIS1 promotes leukaemic homing and engraftment by transactivating SYTL1, which determines intracellular trafficking of CXCR4 and thereby activates CXCL12/CXCR4-dependent migration.
MEIS1-mediated transactivation of synaptotagmin-like 1 promotes CXCL12/CXCR4 signaling and leukemogenesis · Yokoyama T, Nakatake M, Kuwata T, et al. · J Clin Invest 2016What this evidence cannot show
Murine HOXA9/MEIS1 myeloid leukaemogenesis, not a KMT2A fusion model and not B-lineage (rule 73). The mechanism is indirect - MEIS1 does not bind the CXCR4 promoter in this account, it controls a trafficking protein - which matters for the prediction: menin inhibition might reduce CXCR4 FUNCTION without reducing CXCR4 TRANSCRIPT, and a transcriptome-only re-analysis would then return a false negative.
Evidence that argues against this (1)
- This is the strongest internal objection to the re-analysis proposed here, and it is recorded rather than argued away: if the MEIS1 link runs through SYTL1-dependent receptor trafficking rather than through CXCR4 transcription, then RNA-seq is the wrong assay and SYTL1 - not CXCR4 - is the transcript to read out. The re-analysis should therefore score both. Yokoyama T et al. 2016
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In RS4;11 (human MLL-AF4 / KMT2A::AFF1 B-precursor line), separate knockdown of MEIS1, HOXA7, HOXA9 or HOXA10 each hampered the CXCR4/SDF-1 axis - reduced migration toward an SDF-1 gradient and loss of SDF-1-augmented proliferation.
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 abstract does not state whether CXCR4 surface protein or transcript fell, only that the axis was hampered as a function. That ambiguity is exactly the one E4's caveat identifies and is the reason this absence is worth closing with an assay that measures both transcript and surface receptor.
Evidence that argues against this (1)
- Knocking down a transcription factor is not the same perturbation as inhibiting menin: menin loss releases the fusion from a subset of target promoters and the resulting HOXA/MEIS1 collapse is partial and time-dependent, so a knockdown phenotype does not guarantee a drug phenotype at achievable exposures. Ling Q et al. 2023
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
Deposited public human transcriptomes under menin inhibition in KMT2A-rearranged AML exist: a GEO db=gds query returns 88 records for 'revumenib' and 259 for 'menin inhibitor', and GSE281506 is an expression-profiling-by-high-throughput-sequencing series of 42 human samples titled 'Combining Menin and MEK inhibition to target poor prognosis KMT2A-rearranged RAS pathway-mutant acute myeloid leukemia'.
GSE281506 - Combining Menin and MEK inhibition to target poor prognosis KMT2A-rearranged RAS pathway-mutant acute myeloid leukemia (expression profiling by high throughput sequencing, Homo sapiens, 42 samples) · NCBI Gene Expression Omnibus 2026What this evidence cannot show
The GEO free-text counts are loose - 'menin inhibitor' as a db=gds term matches record text broadly and the 259 is an upper bound on relevance, not a count of usable series. GSE281506 itself was confirmed by esummary (accession, gdstype, sample count, taxon and title all read back) but its samples were NOT inspected, so whether it contains a menin-inhibitor-alone arm, and in which cell lines, is unverified here. That check is the first step of the re-analysis, not a result of this entry.
Evidence that argues against this (1)
- This graph has a documented case of GEO metadata lying three separate ways about a SuperSeries - elink reporting no SRA link, a BioProject esearch returning a false zero, and an acc.cgi page that was a reCAPTCHA challenge parsing cleanly. The authoritative route is the SOFT family file. An esummary readback, as used here, is stronger than those three but is still metadata and not data. NCBI Gene Expression Omnibus 2026
re-run twice through pubmed_safe: genuine zero on both runs, non-empty translation, positive control 33 - and the untried symmetric form MEN1[tiab] AND CXCR4[tiab] returns one off-topic record
SURVIVES at full text: 65 records, 50 read, all reviews - a third independent confirmation after two clean abstract-level runs
the axis whose behaviour under menin inhibition has never been reported
zero papers name menin and CXCR4 together in title or abstract
the drug class whose effect on the niche axis is unmeasured
bounds the mechanism's untested pharmacology
the named closing route for that absence