Three pairs of twins who both got leukaemia: the patients who proved the fusion happens before birth
also called Ford 1993 twin study · concordant infant twin leukaemia · intraplacental metastasis hypothesis
Three pairs of infant twins each developed leukaemia, and in each pair both babies' leukaemia cells carried exactly the same, individually unique break in the same gene - a break that was not in their normal cells. Since the same accident cannot happen twice by chance in two different children, the cells must have been shared between them before birth, which is how the world learned that infant leukaemia starts in the womb.
The same thing, for a scientist
Ford and colleagues studied three pairs of monozygotic infant twins with concordant leukaemia and showed that within each pair the twins' leukaemic cells shared a clonal HRX/MLL (now KMT2A) rearrangement that was identical between co-twins, unique to that pair, and absent from constitutional DNA. Because KMT2A breakpoints fall in introns across a region of roughly 10 kb and differ between individual patients, an identical junction in two children cannot plausibly have arisen independently. The result establishes two things at once: that the initiating rearrangement occurs in utero, and that concordance in monozygotic twins is explained by intraplacental transfer of a clone arising in one fetus to the other through the shared monochorionic circulation - the 'intraplacental metastasis' hypothesis - rather than by shared inherited susceptibility. [E1]
Why these six children count as a patient story
They are not a series of presentations, and the paper reports the molecular finding rather than the clinical course: it does not give ages at diagnosis, blood counts, symptoms or outcomes for the individual children 1. Including them here is deliberate, and rests on two arguments.
First, the whole premise of the graph's genesis arm rests on these patients. Every statement in this project that the fusion arises prenatally traces back through this paper and to the neonatal blood-spot work that followed it. A reader should be able to see whose data that was.
Second, a case series is the privacy-preserving way to report a story like this. Three pairs of twins with concordant infant leukaemia is a vanishingly rare configuration, and any individual account would be identifying. Reporting them as the paper does - as three pairs, with the molecular result and no clinical detail - is both what the source discloses and the right level of disclosure.
The argument, in one step1 deeper
KMT2A rearrangements in infant leukaemia occur in introns, clustering within a region of about 10 kb, and individual patients have different breakpoints 1. That variability is the load-bearing fact. It means the exact DNA junction in one child's leukaemia is effectively a fingerprint.
Ford and colleagues found that in each of three twin pairs, both twins' leukaemic cells carried the same rearrangement - clonal, identical within the pair, and not present in their constitutional DNA.
Two children cannot independently acquire the same fingerprint. So either the clone arose once and was shared, or it was inherited - and it cannot have been inherited, because it is absent from the germline. It was shared. Monozygotic twins in a monochorionic placenta have a common circulation, so a clone arising in one fetus can pass to the other. The authors call this unequivocal support for the intraplacental 'metastasis' hypothesis for leukaemia concordance in twins, and it dates the initiating event to intrauterine life.
What it does not show
It does not show that the fusion is sufficient. The companion observation - that some twins with the shared clone never develop leukaemia, and that latency between birth and diagnosis varies - is what drives the two-hit framing used elsewhere in this graph. It also does not establish how common prenatal origin is across infant leukaemia generally; that came later, from neonatal blood spots and cord blood, and this paper is the starting point rather than the proof of prevalence.
It also does not tell you anything about what happens to a child with this disease. That is the gap the rest of this arm exists to fill.
Why this case is in the graph
Because it is the only entry in this arm where the patients changed what the field believed, and because it reframes every other entry here.
If the fusion is present at birth, then a newborn with leukaemia cutis and a normal blood count is not an early diagnosis - the disease is simply at a different point in a course that began months before. The neonate in this arm whose skin lesions appeared at three days old, and the congenital case whose placenta contained leukaemic myeloid cells, are both downstream of the process these twins demonstrated.
It also sets the hardest question in the project. If the initiating lesion happens in utero and most children who carry it never get leukaemia, then what the lab is trying to interrupt is not the translocation. It is whatever comes after it.
| age at dx | infants; individual ages not given in the source available to this entry |
|---|---|
| n patients | 6 |
| series structure | three pairs of monozygotic twins with concordant leukaemia |
| fusion | gene-kmt2a |
| fusion description | clonal, non-constitutional HRX/MLL (KMT2A) rearrangement identical within each twin pair and unique to that pair; partner genes not recorded in the source available to this entry |
| course | not reported in the source available to this entry |
| outcome | not reported in the source available to this entry - this is a molecular study, and the entry deliberately does not fill the gap |
| finding | the initiating KMT2A rearrangement occurs in utero; concordance in monozygotic twins is explained by intraplacental transfer of a clone arising in one fetus, not by inherited susceptibility |
| context statistic | approximately 75% of infant acute leukaemias carry a reciprocal translocation between 11q23 and one of several partner chromosomes |
| source type | case report |
| source type detail | case series embedded in a primary molecular research paper |
| consent note | Published primary research paper. Three pairs of twins with concordant infant leukaemia is an extremely rare configuration, so this entry deliberately reports them only at the level the source does - as three pairs, with the molecular result - and adds no clinical or demographic detail whatever. Reporting them individually would be identifying even if the detail were available. |
| privacy review | Nothing added, inferred or reconstructed. Where the source is silent, this entry records that it is silent rather than supplying a typical-case value. |
presentation
[ "not reported in the source available to this entry" ]
- If the fusion is necessary but not sufficient, then a drug that targets the fusion complex targets the necessary step and leaves the sufficient one untouched - which is compatible with high response rates and low cure rates, the pattern menin inhibitors have shown so far.
- A clone that can persist from fetal life to diagnosis without transforming is, by definition, capable of surviving in a quiescent state. Quiescent clones are the standard mechanism by which targeted therapy produces deep remission followed by relapse.
- This is the finding that most threatens the project's central framing. If the interface between menin and KMT2A is essential to maintenance, these patients show it was not sufficient for initiation, and the graph should be explicit that blocking maintenance and preventing disease are different problems.
- If the clone is present at birth in both twins but only one develops leukaemia in some pairs, what is the second event, and is it detectable in the well twin before disease appears? This is the single most valuable prospective cohort nobody has assembled.
- How long is the latency, and what sets it? Concordant twin leukaemias differ in their interval to diagnosis, and that interval is a direct measurement of how long a KMT2A-rearranged clone can sit without transforming.
- Does the fusion partner determine latency? These three pairs are reported without partner assignment in the abstract; a modern reanalysis of archived twin material with breakpoint sequencing would link partner identity to time-to-diagnosis.
- If the transforming lesion is prenatal and the disease is postnatal, at what point in that interval would a menin inhibitor do anything? The graph's therapeutic entries all assume a patient with established disease; nothing addresses the pre-leukaemic window these patients defined.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In three pairs of monozygotic infant twins with concordant leukaemia, the twins' leukaemic cells each shared a clonal but non-constitutional HRX/MLL (KMT2A) rearrangement that was unique to the pair; since KMT2A rearrangements occur in introns clustering within about 10 kb and individual patients have different breakpoints, this provides evidence that the leukaemogenic event originates in utero and unequivocal support for the intraplacental metastasis hypothesis for leukaemia concordance in twins; approximately 75% of infant acute leukaemias carry a reciprocal translocation between 11q23 and one of several partner chromosomes.
In utero rearrangements in the trithorax-related oncogene in infant leukaemias · Ford AM, Ridge SA, Cabrera ME, Mahmoud H, Steel CM, Chan LC, Greaves M · Nature 1993What this evidence cannot show
Abstract read; the 1993 full text was not accessible to this entry, so the clinical detail of the six children - ages at diagnosis, immunophenotypes, fusion partners, treatment and outcomes - is not recorded here and may or may not be in the paper. Three twin pairs is six children: this is a small series, and it is a series of an extremely unusual configuration. The finding has since been confirmed and extended by an independent line of evidence (retrospective analysis of neonatal blood spots and of cord blood), which is why the confidence is 'established' despite the small n - it does not rest on this paper alone. Methodologically it is 1993 Southern-blot-era breakpoint analysis, not sequencing; the identity of the rearrangements within a pair is shown at the level of restriction fragment junctions.
Evidence that argues against this (1)
- The prenatal-origin finding does not make the fusion sufficient for leukaemia: the two-hit framing, in which a second postnatal event is required, is stated explicitly in the later neonatal-leukaemia literature that cites this work, and the existence of twins carrying a shared clone who do not develop disease is the standing qualification. source 2020
the three concordant twin pairs whose shared, non-constitutional KMT2A breakpoints are the founding observation behind the concordance epidemiology
the twin evidence and the blood-spot evidence are the two independent lines establishing prenatal origin; this entry records the patients behind the first
concordant twin pairs give a direct measurement of the interval between a prenatal rearrangement and clinical disease
- related to →KMT2A / MLL1 — the gene that breaks
clonal non-constitutional KMT2A rearrangement identical within each twin pair; breakpoints fall in introns across about 10 kb and differ between individuals
concordant infant leukaemia; approximately 75% of infant acute leukaemias carry an 11q23 translocation
individual patients have different intronic breakpoints, which is what makes a shared junction between two children evidence of a shared clone
independent evidence of prenatal origin in a single child: leukaemic myeloid cells in the placenta and fetal blood at birth
a KMT2A fusion acquired in an aged, JAK2-mutant marrow, against the prenatal acquisition the twin data establish: the lesion is not exclusively a fetal event