MLL fused to MURINE Af4 in human CD34+ cells — the model that works, and the swap that makes it work
also called MLL-Af4 retroviral human CD34 model
The one engineered system that reliably reproduces this leukaemia in human cells does it by fusing the human MLL gene to the mouse version of the partner. The human version does not make enough virus to do the experiment.
The same thing, for a scientist
Human AF4 constructs produce poor retroviral titre, which is the technical reason human CD34+ transduction models of t(4;11) have historically failed. Lin and colleagues fused MLL to **murine Af4**, highly conserved with human AF4, obtaining high-titre retrovirus and efficient transduction of human CD34+ cells, and generated a model of t(4;11) pro-B ALL that **fully recapitulates the immunophenotypic and molecular aspects of the disease**. The model distinguishes MLL-Af4 from MLL-AF9 by differential genomic target binding and consequent gene-expression patterns, and shows that MLL-Af4 cells can assume a myeloid state under environmental pressure while retaining lymphoid-lineage potential — which the authors connect to t(4;11) patients whose leukaemia evaded CD19-directed therapy by myeloid-lineage switch.
The substitution is not cosmetic and is not obviously harmful
This is an engineered model of KMT2A::AFF1 in which the AFF1 moiety is mouse Aff1, not human AFF1 1. Every claim from this system inherits that substitution. The authors' justification — high conservation between mouse and human AF4 — is reasonable and the model's output is the right disease, which is the strongest empirical argument that the swap does not matter for phenotype.
But this graph's interactome beat catalogues human-AFF1-specific contacts, and the transcript-level work shows KHDRBS/Sam68 proteins translationally repressing the MLL-AF4 message through AU-rich stall sites — a sequence-level regulation that a murine Af4 coding sequence would not reproduce faithfully. Anything about fusion dose, transcript regulation or AFF1-specific protein interfaces should be treated as untested in this model rather than as supported by it.
It is the best available human engineered model for lineage questions
Two of this graph's active questions are answered cleanly here: MLL-Af4 and MLL-AF9 bind different genomic targets and produce different diseases, i.e. the partner is instructive rather than incidental; and lineage switch is an intrinsic capacity of MLL-Af4 cells under environmental pressure rather than a drug-induced novelty 1. The PER-485 relapse line in this beat is the human patient-derived counterpart of that second finding.
| model class | retroviral transduction of human CD34+ cells |
|---|---|
| fusion construct | human MLL fused to MURINE Af4 |
| species chimerism | the 3' moiety is mouse, not human AFF1 |
| reproduces pro b all | true |
| reproduces lineage switch | true |
| endogenous translocation | false |
| physiological dose | false |
| reason for the swap | human AF4 constructs give poor retroviral titre |
| obtainable | constructs by request from the originating laboratory |
- Retroviral overexpression sets fusion dose artificially, and this graph now holds direct evidence that fusion dose is a regulated variable with an RNA-binding-protein brake. Any dose-sensitive conclusion from this model is untested.
- The AFF1 moiety is murine. Interface-directed therapeutics designed against human AFF1 sequence features cannot be validated here.
- The model's lineage plasticity means a CD19-directed agent tested in it will show escape, which is realistic — but it also means a CD19-negative outgrowth in this system is not by itself evidence that the drug caused the switch.
- Does a human-AFF1 construct, once titre is solved by a different vector, give the same disease? If not, the murine substitution is doing work and a body of results needs re-reading.
- Does this model reproduce the fetal gene-expression programme? It depends on the CD34+ donor source, which the abstract does not state, and that single fact would place the model relative to the fetal-liver CRISPR system.
- Is the lineage switch in this model reversible, and does menin inhibition alter its direction?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MLL fused to murine Af4, which is highly conserved with human AF4, produces high-titre retrovirus permitting efficient transduction of human CD34+ cells and generates a model of t(4;11) pro-B acute lymphoblastic leukaemia that fully recapitulates the immunophenotypic and molecular aspects of the disease; MLL-Af4 induces a B-ALL distinct from MLL-AF9 through differential genomic target binding leading to specific gene expression patterns, and MLL-Af4 cells can assume a myeloid state under environmental pressure while retaining lymphoid-lineage potential, an incongruity also observed in t(4;11) patients whose leukaemia evaded CD19-directed therapy by myeloid-lineage switch.
Instructive Role of MLL-Fusion Proteins Revealed by a Model of t(4;11) Pro-B Acute Lymphoblastic Leukemia · Lin S, Luo RT, Ptasinska A, et al. · Cancer Cell 2016What this evidence cannot show
Abstract read; full text not read. The fusion partner is murine Af4, not human AFF1, which is stated in the abstract and is the design's defining feature. The model is retroviral overexpression from a heterologous promoter, so fusion dose is not physiological and the endogenous KMT2A allele is intact. The donor CD34+ cells' source (cord blood, fetal, adult) is not stated in the abstract.
Evidence that argues against this (2)
- KHDRBS/Sam68-family RNA-binding proteins translationally repress the MLL-AF4 transcript through AU-rich stall sites, and synonymous mutations removing those sites restore translation and make an otherwise poorly-transforming construct leukaemogenic in vivo. That makes fusion dose a sequence-encoded regulated variable, so a model using a murine Af4 coding sequence expressed from a retroviral promoter has bypassed the control layer rather than modelling it. Okuda H et al. 2022
- A later human CRISPR-engineered model that creates the endogenous translocation in cord-blood CD34+ cells from a single donor reports that the MLL(intron 9)::AF9 construct, not the AF4 one, formed colonies robustly and engrafted in NSG mice. Two engineered human systems therefore disagree about how readily the index fusion transforms, and the difference tracks the method rather than the cell. Radszuweit P et al. 2026
CD34 selection is that model's input; this entry bounds what the gate does and does not purify
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
human MLL fused to MURINE Af4; recapitulates pro-B ALL
- supports →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
MLL-Af4 cells assume a myeloid state under environmental pressure and retain lymphoid potential
the disease state the model produces