Oncofusion

start / graph / ana-q118-hoxa9-is-the-best-retained-mll1-n-locus-in-all-three-lines-while-menin-leaves-it

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q118-menin-at-the-human-index-fusion

HOXA9 is where menin leaves most completely (share ratio 0.14-0.23 over floor) and MLL1-N-reactive material stays most completely (0.74-1.02 over floor), in all three MLL-rearranged lines - so the menin inhibitor decouples menin from the HOXA9 locus without clearing the MLL1 N-terminus from it

also called HOXA9 retention · the decoupled locus · menin out, MLL1-N in place

At HOXA9 - the gene most people think this leukaemia lives on - the drug takes menin off almost completely while the MLL1 signal stays where it was. Whatever is still sitting there after treatment, it is not being removed by a menin inhibitor.

The same thing, for a scientist

**Across the three MLL-rearranged human lines in GSE127507, HOXA9 is the locus with the largest gap between the two antibodies** [E1]. Floor-normalised within-sample share ratios (VTP/DMSO): | line | fusion | MENIN | MLL1-N | gap | |---|---|---:|---:|---:| | RS4;11 | KMT2A::AFF1 | **0.145** | **0.831** | 5.7x | | MOLM13 | KMT2A::MLLT3 | **0.148** | **0.741** | 5.0x | | ML-2 | KMT2A::AFDN | **0.226** | **1.023** | 4.5x | All three HOXA9 TSS windows are at or above the **99.96th percentile** of their line's MLL1-N DMSO distribution and at the 100.00th in two of three MENIN DMSO distributions, so these are not low-baseline artefacts. The floor is 0.981-1.048 for MLL1-N in all three lines, so **0.74-1.02 means no loss relative to the rest of the genome.** **And it reproduces across species.** Q117 measured murine Hoxa9 at 0.660 raw with a FLAG-tagged fusion and recorded it as one of the best-retained loci in its table - *"the Hoxa cluster is among the best-retained loci... while Meis1 collapses to 0.117"*. **HOXA9 retention with menin loss is therefore now observed in four systems: murine MLL-AF9 with a tagged fusion, and three human MLL-r lines with a shared-epitope antibody.** **The contrast that makes it interpretable is MEIS1 in the same table.** At MEIS1 the menin figure in RS4;11 is 0.010 and the MLL1-N figure is 0.245 - both collapse. At HOXA9 menin collapses and MLL1-N does not. **So the drug's effect at the two canonical HOXA9/MEIS1 targets is not the same effect**, which a cluster-level or programme-level summary would hide. **What this does not say.** It does not say the fusion is still at HOXA9 - that is the attribution problem, and the MLL1-N track is the sum of the fusion and wild-type MLL1. It does not say HOXA9 is still transcribed: the depositing paper reports HOXA9 expression falling under VTP-50469, and occupancy and transcription are different quantities. `concept-occupancy-cannot-separate-eviction-degradation-and-relocation` already holds the general form of this caution in this graph.

Go deeper
Where in the HOXA cluster this is, and what the table can and cannot resolve

The deposited ROSE table carries HOXA9, HOXA7 and HOXA6 as separate 6 kb TSS windows and does NOT carry HOXA10, HOXA5 or the HOXA10-HOXA9 read-through 1. So this is a three-window measurement inside a cluster, not a cluster-level one.

The neighbouring windows do not all behave like HOXA9. In RS4;11, HOXA7's floor-normalised MLL1-N ratio is 0.388 and HOXA6's is 0.085, both from much lower baselines (71.5th and 74.2nd percentile of MLL1-N DMSO) - so HOXA9 is the retained window and its immediate neighbours are not, at least at their TSS. In MOLM13 all three are high-baseline (99.52nd for HOXA7) and HOXA7 gives 0.404 against HOXA9's 0.741.

That pattern - retention concentrated at HOXA9 itself rather than across the cluster - is worth more work than this beat gave it. The genome-wide 2 kb bins for RS4;11 MLL1-N and MENIN are on disk and would give the whole cluster at 2 kb resolution in minutes; it was not done.

Facts and figures
scriptanalyses/q118-menin-at-the-human-index-fusion/roseB.py
statusrun
result summaryHOXA9 TSS window: MENIN 0.145/0.148/0.226 and MLL1-N 0.831/0.741/1.023 over floor in RS4;11/MOLM13/ML-2. Murine FLAG-MLLAF9 Hoxa9 was 0.660 (Q117). MEIS1 in the same table collapses for both antibodies.
cell lineRS4;11, MOLM13, ML-2
genome buildhg19
epigenetic readoutno histone mark measured at HOXA9 in this beat; the H3K79me2 samples GSM3635697/98 (RS4;11) and GSM3635687/88 (MOLM13) exist in the series and were not downloaded
shown for fusionKMT2A::AFF1 (RS4;11), KMT2A::MLLT3 (MOLM13), KMT2A::AFDN (ML-2) - MLL1 N-terminus antibody in all three, so the retained material is not attributed to the fusion
inputs
[
  "GSE127507 ROSE TSS signal table"
]
How this could bite us back
  • HOXA9 is the locus the differentiation block is usually attributed to. If a menin inhibitor removes menin from it and leaves the MLL1 N-terminus in place, the drug may be leaving intact exactly the part of the programme the disease depends on most - which is the failure mode Q117 named in mouse and which is now measured in the human index fusion.
  • A combination rationale follows and cuts both ways: a DOT1L inhibitor or a degrader targeting the retained complex would be indicated, and this beat's DOT1L data at HOXA9 gives 0.053 over floor in RS4;11 - DOT1L leaves HOXA9 almost completely there - but 0.992 in MOLM13 and 0.756 in ML-2, where DOT1L stays. So in the INDEX FUSION line the retained MLL1-N material at HOXA9 is already DOT1L-free under the menin inhibitor alone and adding a DOT1L inhibitor may add nothing at that locus, while in the other two lines it would have something to remove. That is a partner-specific combination argument and it rests on n = 1 per line.
Still open
  1. Is the retained MLL1-N signal at HOXA9 the fusion or wild-type MLL1? A C-terminal MLL1 antibody sees wild-type and not the fusion, and no such sample exists in this series; a tagged fusion in RS4;11 would settle it.
  2. What does H3K79me2 do at HOXA9 in the same cells? GSM3635697/98 (RS4;11) and GSM3635687/88 (MOLM13) are in this series and were not downloaded. If occupancy is retained while H3K79me2 falls, the retained complex is catalytically disconnected rather than functional, which is the hypothesis this graph holds as hyp-the-fusion-stays-on-chromatin-and-stops-working.
  3. Does the retention extend across the HOXA cluster at 2 kb resolution, or is it specific to the HOXA9 window? HOXA7 gives 0.381 in RS4;11 from a lower baseline, which hints at the latter, and the binned genome-wide data already on disk would answer it.
Every source behind this entry (1)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Floor-normalised within-sample share ratios at the HOXA9 TSS window under 330 nM VTP-50469 are MENIN 0.145 / MLL1-N 0.831 in RS4;11, MENIN 0.148 / MLL1-N 0.741 in MOLM13, and MENIN 0.226 / MLL1-N 1.023 in ML-2, with all baselines at or above the 99.96th percentile; the deposited table carries HOXA9, HOXA7 and HOXA6 but not HOXA10 or HOXA5.

    What this evidence cannot show

    NEITHER ANTIBODY IN THIS SERIES IS FUSION-SPECIFIC AND THE ANTIBODY NAME SAYS SO. MLL1n is anti-MLL1 N-terminus (Bethyl A300-086A, lots 5 or 6); the N-terminus is retained in KMT2A::AFF1, so that track is the sum of the fusion and wild-type MLL1 in unknown proportion. MENIN is Bethyl A300-105A Lot 8; menin binds the fusion, wild-type MLL1 and non-KMT2A partners. Q117 could attribute its numbers because its fusion was FLAG-tagged; nothing here can. Every occupancy statement in this beat is about MLL1-N-reactive material or menin-occupied sites, never about the fusion. No scale factor exists for any sample. The depositors state that a Drosophila spike-in was run on one batch, that fly/human ratios were 'found to be small', and that NO spike-in normalisation was applied; no number for that measurement is reported and none is deposited. Every cross-condition figure here is therefore a WITHIN-SAMPLE share or a within-sample rank, which cancel an unknown per-sample scale factor exactly, and never a fold change. n = 1 per antibody per condition per line - there is no replicate anywhere in GSE127507 - so no number here carries a spread and none is precise. This is a 6 kb TSS window, not the HOXA cluster and not the gene body. The floor is the aggregate share ratio over ~8,600 windows in DMSO percentile 50-90 and is 0.981-1.048 for the three MLL1-N tracks. MOLM13's MENIN pair is a 4-day exposure while its MLL1n pair is 3 days, so the MOLM13 gap partly reflects one extra day of drug. Occupancy is not transcription: the depositing paper reports HOXA9 expression falling under this drug.

    Evidence that argues against this (2)
    • HOXA9 is among the most CpG-island-dense and most highly occupied windows in every one of these tracks, so a share statistic there is close to its ceiling and a retained high rank is the least surprising outcome. The MEIS1 window in the same table has a comparable baseline and collapses, which is the reason the contrast is reported - but a ceiling argument cannot be fully excluded from a rank or share statistic alone. Iqbal Lab oncofusion knowledge graph 2026
    • The depositing paper's own account is that VTP-50469 displaces the fusion from a SUBSET of its binding sites and that HOXA9/MEIS1 expression falls. A reader could take HOXA9 as being in the non-displaced subset, in which case this entry adds a number to a known result rather than a new phenomenon - and that is a fair reading of the MLL1-N half. The new part is the 5-fold menin/MLL1-N gap at the same window. Krivtsov AV et al. 2019
Connected to