KMT2A::MLLT11 (MLL-AF1q), t(1;11)(q21;q23) — 28 patients, and not one functional experiment

From Table 1 of the 2023 recombinome, read directly from the full text rather than from a summary 1:

28 patients, rank 10 of all direct KMT2A partners, and 2.4% of the 1116 AML patients in the series. The founding cases were infants with acute myelomonocytic leukaemia 2; the 1999 series of six cases confirmed the specific association of AF1q with t(1;11)(q21;q23)-positive acute leukaemia with monocytic involvement 3.

The lineage skew is the sharpest of any partner in this beat. EPS15, at a comparable patient count, is predominantly an infant ALL partner; MLLT11 at 28 patients has one lymphoid case in the whole series. Any mechanistic account of this fusion has to explain a monocytic/myeloid bias, and none has been attempted.

Counted-absence claims in this project are only as good as the symbol set behind them, so the searches are given in full and were run in both nomenclature directions, current and retired 5:

• (MLL OR KMT2A) AND (MLLT11 OR AF1q) [Title/Abstract] → 20 records, all titles read.
• AF1q OR MLLT11 [Title/Abstract] → 88 records, all titles read. The great majority are wild-type AF1q in solid tumours (breast, colorectal, ovarian, gastric, oesophageal, neuroblastoma, osteosarcoma, glioma) or Mllt11 in mouse neurodevelopment.
• "MLL-AF1q" OR "MLL/AF1q" OR "MLL-MLLT11" OR "KMT2A-MLLT11" [All Fields] → 10 records, all titles read; all are case reports, cytogenetic cohort descriptions, or multiplex RT-PCR fusion-panel method papers.
• The same fusion terms AND (mice OR mouse OR retrovir* OR colony OR replating) → 0 records.
• (MLL OR KMT2A) AND (MLLT11 OR AF1q) AND (transform* OR immortali* OR leukemogen*) → 4 records, all titles read; none is a functional study of the fusion.

The retired symbol is the one that retrieves the literature here, as the septin work predicted: AF1q returns 58 records against MLLT11's 45, and the KMT2A intersection is dominated by AF1q. But the opposite failure mode does not bite: AF1q is not a common word and is not an acronym another field uses, and every count above was title-checked, so no inflation correction was needed.

What this bounds: there is no published experiment in which the KMT2A::MLLT11 protein was expressed and a phenotype measured. What it does not bound: a negative or unpublished result, or a construct buried in a supplementary table of a screen. Both are possible and neither is retrievable.

The dispatch asked whether MLLT11 transforming would break the elongation-hijack model and the forced-oligomerisation model. The honest answer is that nobody knows whether it transforms, and that is a more interesting result than either outcome would have been, for three reasons.

1. The models have never been exposed to their hardest case. Both models were built on partners chosen because they had a module to point at — a YEATS domain, a coiled coil, a 15-residue oligomerisation segment. MLLT11 has no module to point at and is the 10th most common partner in the disease. That it has never been tested is a statement about how the field chose its experiments, not about the biology.
2. A negative result would be as informative as a positive one. If KMT2A::MLLT11 does not transform mouse progenitors in a standard assay, then a fusion found in 28 patients is not sufficient for leukaemia on its own, which would force a cooperating-lesion account — and there is a hint in that direction, since the one reported three-way t(1;11) case carries MYO18A as a reciprocal partner and the 1999 series included a three-way translocation 37.
3. The oligomerisation question is decidable with existing reagents. Refolded recombinant AF1q exists and has been through analytical ultracentrifugation; the published abstract simply does not report its oligomeric state 6. One sedimentation-velocity run would establish whether the forced-oligomerisation model has a real counterexample or a false one.

So the model-level conclusion is: forced oligomerisation and elongation hijack both remain intact, not because MLLT11 supports them, but because MLLT11 has never been allowed to argue.

Neutral, and untested — but structurally favourable in one specific respect worth recording.

Every reported KMT2A::MLLT11 junction lies at or beyond KMT2A exon 7 34, so the retained KMT2A moiety includes the extreme N-terminus where MBM1/MBM2 sit. If menin dependence is a property of that retained N-terminus, KMT2A::MLLT11 should be menin-inhibitor-sensitive regardless of the partner contributing nothing. Nobody has tested it, and there is no cell line, so the test needs a construct made first.

Nothing in the MLLT11 literature threatens the menin hypothesis. The one place it could have — an alternative, menin-independent transforming mechanism supplied by the partner — is precisely the thing that has never been looked for.

Cell type. Patient-level only: infant and paediatric AML with monocytic involvement 13. No experimental cell-of-origin work exists for this fusion.

Cell state. Not characterised for the fusion. Wild-type AF1q is a progenitor-restricted, Notch-responsive lineage-decision protein that pushes multipotent progenitors toward T and away from B 6 — which runs opposite to the fusion's myeloid clinical phenotype, an unexplained tension recorded in gene-mllt11.

Subcellular and subnuclear localization. Not reported for the fusion protein, at all. Wild-type AF1q shuttles, with pronounced karyophily, constitutive nuclear export and pericentrosomal ubiquitin-dependent degradation 6. The fusion retains the complete AF1q ORF and therefore the annotated nuclear export signal at aa 24-32 — which would make it the only KMT2A fusion known to carry a functional NES, if the NES is functional in that context. Nobody has imaged it.

Epigenetics. Nothing. No H3K79me2, no H3K4 methylation, no HOXA-locus occupancy, no transcriptome, for the fusion or for any patient sample carrying it.