Reciprocal status is a different fact depending on whether you assay DNA or RNA, and the whole prognostic literature is RNA
also called reciprocal transcript versus allele · 45 versus 80 percent · reciprocal status assay dependence
Whether a patient counts as having the mirror-image fusion depends on what you measure. Look at the DNA and almost everyone has it; look at the RNA and fewer than half do. Everything we know about whether it matters comes from the RNA test.
The same thing, for a scientist
In t(4;11), about 80% of patients carry both reciprocal fusion alleles at the genomic level and the remaining 20% carry a reciprocal KMT2A allele fused to some other partner or a solitary 3'-KMT2A segment [E1]; across the whole KMT2A recombinome, 94% of rearrangements generate a reciprocal fusion gene [E2]. But only 45% express a detectable reciprocal transcript [E3]. Three things could explain the gap, and they have different clinical meanings: (i) the reciprocal allele is intact but transcriptionally silent; (ii) the join is out of frame or to an unexpected partner, so an AFF1-junction RT-PCR cannot see it even when the allele is transcribed - the genomic study found exactly this, with reciprocal KMT2A fusions to NFKB1 and RABGAP1L actively transcribed [E1]; (iii) the allele was lost by clonal selection [E4]. The graph should therefore stop writing 'reciprocal-positive' without saying which assay, because the prognostic result and the mechanistic literature are on opposite sides of this line: the survival split is RT-PCR [E3], the allele counts are DNA [E1][E2], and the mechanistic models are transfected constructs.
The assay-dependence is not a technicality, it is the finding
An RT-PCR designed around the AFF1 exon 3 / KMT2A exon 12 junction is blind by construction to a reciprocal allele in which 3'-KMT2A is fused to NFKB1, RABGAP1L, PBX1, ELF2, DSCAML1 or FXYD6 - all of which were found in t(4;11) patients scored reciprocal-negative 1. So the 45% figure is a lower bound on reciprocal expression, by an amount nobody has measured. If the reciprocal-negative arm of the prognostic analysis is contaminated with patients who do express a reciprocal KMT2A product through an unexpected partner, the true effect size is larger than reported, not smaller - which strengthens rather than weakens the clinical observation, while making its mechanistic interpretation harder.
What an honest per-patient statement looks like
'Reciprocal AFF1::KMT2A transcript detected / not detected by RT-PCR at diagnosis' - which is all any published cohort measures. Not 'the patient has / lacks the reciprocal fusion'. And never 'the patient has / lacks the reciprocal protein', because no protein-level assay exists that distinguishes the reciprocal-derived MLL-C fragment from the wild-type one 5.
| which assay underlies the prognostic result | RT-PCR transcript detection |
|---|
assay dependence
{
"dna_t411_both_alleles_pct": 80,
"dna_any_partner_reciprocal_pct": 94,
"rna_t411_transcript_pct": 45,
"protein": "no discriminating assay exists"
}explanations for the gap
[ "transcriptional silencing", "join to an unexpected partner invisible to junction RT-PCR", "out-of-frame join", "clonal loss" ]
- A biomarker that decides therapy on reciprocal status would, as currently assayed, misclassify an unknown fraction of patients whose reciprocal allele is fused to a partner the primers cannot see.
- No protein-level assay exists, so a patient could be transcript-positive and protein-negative and there is no way to tell.
- How many patients scored reciprocal-negative by junction RT-PCR carry a transcribed reciprocal KMT2A allele fused to another partner? Genomic sequencing of the reciprocal-negative arm of any existing cohort would answer it.
- Is the prognostic effect larger when reciprocal status is scored by DNA rather than RNA? If the negative arm is contaminated, it must be.
- Can a protein-level assay for the reciprocal product be built at all? It would need an epitope spanning the AFF1-KMT2A junction, which no commercial reagent provides.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In t(4;11) patients scored as lacking the reciprocal allele, genomic analysis found cryptic and three-way translocations, in-frame reciprocal MLL fusions to novel partners such as NFKB1 and RABGAP1L that were actively transcribed in leukaemic cells, out-of-frame reciprocal fusions to PBX1, ELF2, DSCAML1 and FXYD6, and patients carrying only solitary 3'-MLL segments.
Complex MLL rearrangements in t(4;11) leukemia patients with absent AF4.MLL fusion allele. · Kowarz E, Burmeister T, Lo Nigro L, et al. · Leukemia 2007What this evidence cannot show
Selected subgroup rather than consecutive series; n not stated in the abstract; the categories are illustrative rather than exhaustively counted.
Evidence that argues against this (1)
- The clinical literature has not adopted this: the cohort study that established the prognostic effect scores reciprocal status by a single RT-PCR and reports a binary, so the heterogeneity described here is absent from every outcome analysis. Agraz-Doblas A et al. 2019
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
An additional reciprocal fusion gene is generated in 94% of KMT2A rearrangements in the Frankfurt DCAL series, and 114 of 2,454 patients carried the reciprocal allele alone.
The reciprocal world of MLL fusions: A personal view. · Marschalek R · Biochim Biophys Acta Gene Regul Mech 2020What this evidence cannot show
Single-centre referral series presented as a personal view; allele detection reflects that centre's assay panel.
Evidence that argues against this (1)
- Allele generation is not expression: fewer than half of t(4;11) patients express a detectable reciprocal transcript. Agraz-Doblas A et al. 2019
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The reciprocal transcript was detected in 45% (19/43) of t(4;11) cases by RT-PCR in the Interfant cohort, and HOXA cluster genes were expressed exclusively in the reciprocal-expressing patients.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al. · Haematologica 2019What this evidence cannot show
Diagnostic material only, one timepoint, one assay; the HOXA association means reciprocal status is confounded with a transcriptional subgroup from the outset.
Evidence that argues against this (1)
- An RT-PCR keyed to the AFF1 junction cannot detect a reciprocal KMT2A allele fused to a different partner, which genomic analysis shows occurs and is transcribed. Kowarz E et al. 2007
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Loss of the AF4-MLL allele in about 40% of patients is interpreted by the Frankfurt group as a clonal evolutionary process that selects for a more aggressive leukaemia.
The role of reciprocal fusions in MLL-r acute leukemia: studying the chromosomal translocation t(4;11). · Wilhelm A, Marschalek R · Oncogene 2021What this evidence cannot show
An interpretation offered in a modelling paper's framing, not a measurement; no paired diagnosis/relapse data supports the word 'loss'.
Evidence that argues against this (1)
- The cross-sectional data cited for this interpretation reports diagnostic expression status and makes no claim about change over time. Agraz-Doblas A et al. 2019
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In t(4;11) lines, the cleaved AF4-MLL species was identified in immunoprecipitations as a band at approximately 194 kDa, with the assignment made on apparent molecular weight in a complex also containing RUNX1, MLL-C and wild-type AF4.
RUNX1 is a key target in t(4;11) leukemias that contributes to gene activation through an AF4-MLL complex interaction. · Wilkinson AC, Ballabio E, Geng H, et al. · Cell Rep 2013What this evidence cannot show
The reagent class used cannot distinguish the reciprocal-derived MLL-C fragment from the wild-type one, because Taspase1 releases sequence-identical fragments from both; the molecular-weight assignment is the only discriminator offered.
Evidence that argues against this (1)
- An MLL1 antibody specific for the 180 kDa MLL1 C-terminal fragment does not detect the MLL-AF9 fusion, a property that study relied on - showing the reagent class can separate wild-type from a DIRECT fusion, which is precisely the case that does not apply to a reciprocal fusion that retains the same C-terminus. Xu J et al. 2016
Explains why the allele count and the transcript count differ by a factor of two.
Extends the reagent problem from immunoprecipitation to the clinical assay layer: DNA, RNA and protein give three different answers.
Specifies which assay that proposed reanalysis would have to use, and what it would miss.
Whether loss has occurred cannot be separated from which assay was used.
Adds a fourth layer to the assay problem: a patient can be junction-negative and still make a menin-blind KMT2A protein.