Oncofusion

start / graph / mechanism-in-utero-origin-neonatal-blood-spot-backtracking

Mechanismestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q1b-genesis2

In-utero origin, proven: clonotypic KMT2A::AFF1 junctions found in the patient's own newborn blood spot

also called Guthrie card backtracking · neonatal blood spot backtracking · prenatal origin of infant leukaemia

Every baby in most countries has a few drops of blood dried onto a card in the first days of life. Researchers went back to those cards for children who later developed leukaemia and found the child's exact leukaemia fusion sequence already there at birth, so the first event happened before the baby was born.

The same thing, for a scientist

The prenatal origin of KMT2A-rearranged infant leukaemia is not an inference; it is a direct observation. Gale et al. developed a nested-PCR assay to interrogate archived neonatal Guthrie cards using patient-specific genomic KMT2A::AFF1 breakpoint sequences as clonal fingerprints, and recovered those unique junctions from the birth blood of children later diagnosed at 5 months to 2 years [E1]. Because a genomic breakpoint junction is effectively unique to one clone (the exact nucleotide position and the microhomology or insertion at the join differ between patients), a positive backtrack is not an epidemiological association but an identity match. The same method later demonstrated prenatal origin for ETV6::RUNX1 and, more surprisingly, for RUNX1::RUNX1T1 in children diagnosed as old as 10-14 years, establishing that a translocated clone can persist for over a decade before overt disease [E2][E3]. For KMT2A the interval is short, which is the key point of contrast: the fusion appears to be much closer to sufficient here than in any other backtracked subtype.

Go deeper
Why a genomic breakpoint is a fingerprint and an RNA transcript is not1 deeper

Backtracking works because the assay targets genomic DNA across the breakpoint, not the fusion mRNA. Two patients with the same cytogenetic t(4;11) almost never share the same nucleotide-level join: the break falls at a different position inside the KMT2A breakpoint cluster region, and error-prone end joining leaves patient-specific microhomologies, deletions or filler sequence at the junction 1. A patient-specific PCR therefore cannot amplify anyone else's clone, which removes the contamination objection that would sink an RT-PCR-based screen. This distinction matters directly for the newborn-screening question (see target-newborn-screening-fusion-transcripts): backtracking is retrospective and patient-specific, whereas screening would have to be prospective and generic, a much harder and much less specific assay.

Sensitivity, and what a negative backtrack does not mean

A Guthrie punch samples only a few microlitres of the newborn's blood, so a negative backtrack in a genuinely fusion-positive newborn is entirely expected if the clone sits below the assay's detection limit. Published detection rates are therefore lower bounds on the true prenatal-origin fraction, never estimates of it 13. Every quantitative statement in this beat that rests on backtracking inherits that one-sided error.

Protracted latency in other fusions, short latency here

Maia et al. extended backtracking to KMT2A-fusion-positive and ETV6::RUNX1 cases diagnosed outside the usual age range and showed the preleukaemic state can occasionally persist for many years 3. Wiemels et al. found clonotypic RUNX1::RUNX1T1 in the birth spots of children diagnosed more than ten years later and, critically, found the same genomic fusion sequence still present in three patients who were in complete clinical remission, i.e. a fusion-positive but non-leukaemic clone coexisting with health 2. Infant KMT2A-r leukaemia sits at the opposite end of that spectrum, with a latency measured in months (see epi-latency-in-utero-to-diagnosis-kmt2a).

How this could bite us back
  • Proving prenatal origin makes primary prevention an obstetric problem, not a paediatric one, and the intervention window closes at birth. Any prevention strategy that starts after delivery is aimed at the wrong event.
  • If a fusion-positive clone can sit in complete remission without causing disease [E2], then 'fusion detected' is not interchangeable with 'disease present', and a fusion-transcript MRD assay may over-call relapse risk in exactly the patients a menin inhibitor is being used to keep in remission.
Still open
  1. Nobody has backtracked a KMT2A::AFF1 case with modern deep sequencing rather than nested PCR. Targeted capture of the KMT2A breakpoint cluster region from a whole Guthrie card would convert a set of anecdotes into a detection-rate estimate with a real denominator.
  2. Has anyone looked for the RECIPROCAL AFF1::KMT2A junction in birth spots? If the reciprocal is present at birth and lost by diagnosis, that dates reciprocal loss to postnatal life and makes it a candidate second event in a disease with almost no mutational second events.
  3. Is the fusion at birth already confined to a single cell type, or present across lineages? Sorting is impossible from a dried spot, so this needs a prospective cord-blood cohort, but the answer would say whether the target cell is already committed in utero.
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Patient-specific genomic KMT2A::AFF1 fusion junctions are detectable in the archived neonatal blood spots of children subsequently diagnosed with infant ALL, proving the translocation arose in utero.

    Backtracking leukemia to birth: identification of clonotypic gene fusion sequences in neonatal blood spots · Gale KB, Ford AM, Repp R, Borkhardt A, Keller C, Eden OB, Greaves MF · Proceedings of the National Academy of Sciences of the USA 1997
    PMID 9391133DOI
    What this evidence cannot show

    Small number of cases; the assay is one-sided, because a negative result cannot exclude prenatal origin when the Guthrie punch samples only microlitres of blood. Published before modern anti-contamination practice, though the use of patient-unique junctions largely protects against cross-contamination by design.

    Evidence that argues against this (2)
    • The detection rate is low enough that backtracking is close to uninformative for MLL fusions outside infancy. Burjanivova et al. ran the identical patient-specific approach on the largest reported cohort of non-infant childhood leukaemia Guthrie cards, with an assay calibrated by serial dilution to detect the preleukaemic clone if 1-3 positive cells were present in the spot. They recovered the clonotypic marker in 3 of 12 ALL patients (25%) and in 0 of 13 AML patients -- including the three cases whose marker WAS an MLL fusion (MLL/AF6, MLL/AF9, MLL/AF10). So for MLL-fusion leukaemia specifically, the published backtracking record outside infancy is zero for three. The entry already says a negative backtrack is one-sided; this quantifies how one-sided, and it means no prevalence, penetrance or latency statement anywhere in this beat can be derived from backtracking frequencies. Read: abstract only. Burjanivova T et al. 2006
    • A PCR positive for a leukaemia fusion is not, on its own, evidence of a clone destined for disease. Song et al. applied sensitive nested RT-PCR to a large series of healthy peripheral blood samples and detected every marker tested -- MLL partial tandem duplication, BCR-ABL p190 and p210, MLL-AF4, AML1-ETO, PML-RARA and CBFB-MYH11 -- in healthy individuals at measurable prevalence, and also found aberrations in cord blood from healthy neonates. Their conclusion was that these lesions are not themselves sufficient for transformation. This does NOT undercut Gale's specific logic, because a patient-unique genomic breakpoint junction is a different and far more specific assay than a generic RT-PCR; the entry is right that the two are not interchangeable. It does undercut the looser reading that 'the fusion was there at birth' is the same statement as 'the leukaemia had already begun'. Read: abstract only (PMC full text returned empty, so the per-marker prevalence figures were not obtained). Song J et al. 2011
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Clonotypic fusion sequences can be present at birth, persist for more than a decade, and remain detectable in patients who are in complete clinical remission.

    In utero origin of t(8;21) AML1-ETO translocations in childhood acute myeloid leukemia · Wiemels JL, Xiao Z, Buffler PA, Maia AT, Ma X, Dicks BM, et al. · Blood 2002
    PMID 11986239DOI
    What this evidence cannot show

    Different fusion (RUNX1::RUNX1T1) and different disease; 5 of 10 cases positive; persistence in remission assessed in only 3 patients. Whether the same holds for KMT2A::AFF1 has not been tested.

    Evidence that argues against this (1)
    • An alternative explanation for 'fusion still detectable in complete remission' is assay background rather than a persisting preleukaemic clone. Song et al. detected AML1-ETO and other leukaemia-associated fusion transcripts by nested RT-PCR in healthy individuals with no leukaemia history, at prevalences they interpreted as showing such lesions are not sufficient for transformation. Wiemels' remission-positive finding in three patients is therefore compatible with two readings -- a surviving preleukaemic clone, or the same background positivity seen in healthy donors -- and the published work does not separate them. The bite-back drawn from E2 about fusion-transcript MRD over-calling relapse risk is strengthened by this, not weakened. Read: abstract only. Song J et al. 2011
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    The postnatal preleukaemic interval after a prenatal KMT2A or ETV6::RUNX1 fusion can occasionally be very protracted.

    Protracted postnatal natural histories in childhood leukemia · Maia AT, Koechling J, Corbett R, Metzler M, Wiemels JL, Greaves M · Genes, Chromosomes & Cancer 2004
    PMID 14978794DOI
    What this evidence cannot show

    Explicitly a study of rare cases selected for being diagnosed outside the normal age range, so it is an existence proof, not a frequency estimate. Selection for unusual cases is both the point of the design and its limit.

    Evidence that argues against this (1)
    • Prenatal presence of the fusion is necessary but demonstrably not sufficient, which limits how much the in-utero origin result can carry. Montes et al. expressed MLL-AF4 in human cord-blood CD34+ HSPCs, the closest available model of the cell that carries the fusion at birth: it enhanced hematopoietic repopulating function and clonogenic potential but did not immortalise or transform, and adding FLT3 tyrosine-kinase-domain mutation or FLT3 overexpression -- the most obvious candidate cooperating lesion given the enormous FLT3 levels in MLL-AF4 pro-B-ALL -- still produced no disease in any transplanted animal over 16 weeks. The protracted-latency literature is therefore not only about rare selected cases; it is consistent with a general requirement for something the fusion does not supply. Read: abstract only. Montes R et al. 2014
Connected to