Lineage switch to myeloid leukaemia under CD19-directed therapy in ZNF384-rearranged ALL - and in the one cohort that caught both families, the ZNF384 switch looked mutationally NOTHING like the KMT2A ones
also called ZNF384 lineage switch · TCF3::ZNF384 CAR-T lineage switch · blinatumomab lineage conversion ZNF384
Leukaemias with a ZNF384 fusion can escape CD19-targeted drugs by turning into myeloid leukaemia, the same escape route already known for KMT2A. But in the one study that caught both kinds in the same group of children, the KMT2A escapes carried almost no new mutations while the ZNF384 escape carried a large set - so the two families may be doing the same thing by different means.
The same thing, for a scientist
Among **182 children with BCP-ALL treated with blinatumomab**, six lineage switches occurred - **17.4% of all resistance cases (4/27) and 3.2% of relapses (2/63)**. **Five carried KMT2A rearrangements and one carried TCF3::ZNF384** [E1]. That single cohort is, as far as this beat could find, the only place the two fusion families' lineage switches have been observed side by side under the same therapy, and the comparison produced a divergence nobody predicted: verbatim, *"KMT2A-r patients demonstrated very few additional mutations, while in the TCF3::ZNF384 case, lineage switch was accompanied by a large set of additional mutations"* [E1]. The authors' conclusion is the generalising one: *"lineage switch is not restricted to KMT2A-rearranged cases and can be anticipated in ALL with ZNF384 rearrangement"*, with an explicit monitoring recommendation naming both [E1]. Independently, myeloid lineage switch after **CD19 CAR-T** has been reported in a TCF3-ZNF384-positive patient [E2], and the Ponte di Legno group advise that anti-CD19 CAR-T *"should be used carefully, if at all, in patients with ZNF384 rearrangements"* [E3]. A pre-immunotherapy case exists too: lineage switch from pro-B ALL to AML with a TAF15-ZNF384 rearrangement, reported in 2013 [E4]. **The mechanistic prior is strong before any therapy is given: the founding lesion sits in a subset of HSCs, ZNF384 rearrangement accounts for about half of paediatric B/myeloid MPAL, and 78-100% of these cases already express CD33.**
The one head-to-head observation, and what it does and does not license
The Russian blinatumomab series is the comparison this graph has been missing, because both families appear in one cohort under one therapy with one molecular workup 1.
| KMT2A-rearranged (n = 5) | TCF3::ZNF384 (n = 1) | |
|---|---|---|
| additional mutations at switch | "very few"; only one of five acquired pathogenic events in AML relapse | "a large set of additional mutations"; TP53 p.R89Q and TP53 p.R16H recorded |
| karyotype at diagnosis | t(4;11), t(11;19) and complex | normal, 46,XX - the fusion was cryptic, found by FISH, RNA-seq and exome |
| switch pattern | complete conversion to CD19-negative AML in several | retained a CD19-positive B population and acquired a CD19-negative myeloid one; scored as ALAL (B + myeloid) |
| clonal architecture | clone loss plus new clones in one; stable IG/TR in two of six overall | "Leukemic clone persistence" |
What this licenses: ZNF384-rearranged disease belongs on the lineage-switch watch list, which the authors say explicitly. And the two families may reach the same clinical endpoint by different routes - a mutation-poor chromatin reprogramming in KMT2A, a mutation-accompanied clonal shift in ZNF384.
What it does not license, and the reason this section exists: n = 1 for the ZNF384 arm. One patient with many mutations against five with few is not a difference between families; it is one observation. This project has been caught before turning a single case into a between-group claim, and the honest statement is that the contrast is a hypothesis generated by one cohort, with the mechanism recorded in mechanism-lineage-switch-kmt2a - where the KMT2A literature already establishes that BOTH subclone selection and direct reprogramming occur, in roughly equal measure - as the place a proper test would start.
Why the prior was high before any of these cases
Three facts about ZNF384-rearranged leukaemia make lineage escape from a B-lineage-directed agent the expected failure mode rather than a surprise:
- The founding lesion is in a stem cell. TCF3-ZNF384 was detected in purified blasts, in multiple myeloid- and lymphoid-potent progenitors and in a subset of HSCs - and the same paper contrasts this with Ph-like B-ALL, where founding lesions stop short of the HSC compartment 5. A reservoir above lineage commitment is a reservoir a lineage-directed drug cannot reach.
- The disease is already half myeloid. ZNF384 rearrangement accounts for 48% of paediatric B/myeloid MPAL, and ZNF384-rearranged MPAL and ZNF384-rearranged B-ALL have indistinguishable gene-expression profiles 5 - so the B-ALL presentation is a phenotypic position on a continuum, not a fixed identity.
- CD33 is already on. 78-100% of ZNF384-rearranged cases express CD33 and 38-100% express CD13 3. The myeloid programme does not need to be created under pressure; it needs to be selected.
The corollary for monitoring is the authors' own: "careful investigation of all BM compartments prior to and after course of blinatumomab, especially in patients with KMT2A rearrangements and ZNF384 fusions, can result in timely diagnosis of lineage switch", with CD45 and light-scatter properties as the practical handle since CD45 is already in every MRD panel 1.
| kind | resistance |
|---|---|
| severity | high - conversion to CD19-negative myeloid leukaemia removes the target entirely and is a recognised cause of treatment failure |
| mitigation | Monitor all bone-marrow compartments before and after CD19-directed therapy, using CD45 and light-scatter properties which are already in standard MRD panels, in patients with ZNF384 fusions as well as those with KMT2A rearrangements; do not rely on a CD19-gated MRD assay alone, because the escaping population is CD19-negative by definition. |
affects
[ "drug-blinatumomab", "CD19-directed CAR-T (no drug node in this graph)" ]
measurements — the numbers behind the chart
[
{
"quantity": "proportion",
"subject": "lineage conversion among blinatumomab resistance cases in paediatric BCP-ALL",
"value": 0.174,
"units": "fraction",
"n": 27,
"universe": "27 resistance cases among 182 paediatric BCP-ALL patients treated with blinatumomab; 4 of 27 converted",
"null": null,
"note": "NOT a ZNF384-specific rate - this is all BCP-ALL, and only one of the six switches in the whole series was ZNF384-rearranged. The number of ZNF384-rearranged patients in the 182 is not reported, so no subtype rate can be derived."
},
{
"quantity": "proportion",
"subject": "lineage conversion among blinatumomab relapses in paediatric BCP-ALL",
"value": 0.032,
"units": "fraction",
"n": 63,
"universe": "63 relapses among 182 paediatric BCP-ALL patients treated with blinatumomab; 2 of 63 converted",
"null": 0.174,
"note": "the null is the resistance-case rate in the same cohort - conversion is five-fold commoner among resistant patients than among relapsing ones, which is a statement about when the escape happens"
},
{
"quantity": "count",
"subject": "ZNF384-rearranged lineage switches in the only cohort containing both fusion families",
"value": 1,
"units": "patients",
"n": 6,
"universe": "six lineage switches among 182 blinatumomab-treated paediatric BCP-ALL patients; the other five were KMT2A-rearranged",
"null": 5,
"note": "THE NULL IS THE KMT2A COUNT IN THE SAME SERIES. n = 1 is the reason the mutation-burden contrast between the families is an observation and not a finding."
},
{
"quantity": "proportion",
"subject": "CD33 expression in ZNF384-rearranged childhood B-ALL",
"value": 0.78,
"units": "fraction of cases, lower bound of the reported range",
"n": 218,
"universe": "Ponte di Legno series; CD33 positive in 78-100% of cases across partner subgroups, CD13 in 38-100%",
"null": null,
"note": "recorded as the prior: the myeloid programme is already expressed before any therapy, so a myeloid escape needs selection rather than creation"
}
]- This entry IS a bite-back, against CD19-directed therapy in a subtype that is half myeloid before treatment starts. The specific failure mode is that the target antigen is not lost by mutation but by lineage, so no CD19 sequencing at relapse will explain it and no CD19-gated MRD assay will see it coming.
- The mitigation carries its own cost. Monitoring all marrow compartments and abandoning CD19-gated MRD adds complexity to every ZNF384-rearranged patient's follow-up on the basis of a risk nobody has quantified, and over-calling a small myeloid population as an incipient switch would push patients toward AML-directed therapy - which the paediatric MPAL literature in this graph says is the worse choice, with ALL-type induction giving 80% 5-year EFS against 36% for AML-type.
- And withholding CD19-directed therapy has a mortality cost of its own in relapsed disease. 'Carefully, if at all' rests on one case report plus one case in a 182-patient series; the balance of harms has not been computed by anyone and should not be presented as though it had.
- Does the mutation-burden contrast between KMT2A and ZNF384 lineage switches hold beyond one patient? If KMT2A switches are mutation-poor chromatin reprogramming and ZNF384 switches are mutation-accompanied clonal shifts, the two families would need different monitoring and different salvage logic. The test is a pooled reanalysis of paired diagnosis/switch exomes across published lineage-switch cases, stratified by fusion family - the data exist in several series.
- What is the lineage-switch rate in ZNF384-rearranged patients receiving CD19-directed therapy? The Ponte di Legno recommendation is 'carefully, if at all', and nobody has a denominator. Until someone does, a precautionary recommendation is being made against one of the most effective salvage options in relapsed B-ALL.
- Is CD33-directed therapy a better or worse idea here? CD33 is positive in 78-100% of cases, which makes gemtuzumab-class agents superficially attractive, but the same plasticity that produces CD33 can remove it - and a CD33-negative B-lymphoid escape is exactly as available as the CD19-negative myeloid one. Nobody has tested a myeloid-directed agent in this subtype.
- Is the CD19-positive-plus-CD19-negative pattern seen in the ZNF384 patient - retention of the B population alongside acquisition of a myeloid one - characteristic of this family? It is a different pattern from complete conversion and would imply the escape is expansion of a pre-existing subpopulation rather than conversion of the bulk. The paper notes a small myeloid subpopulation was visible throughout relapse treatment before the switch, which supports that reading in this one case.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Among 182 paediatric BCP-ALL patients treated with blinatumomab, six lineage switches occurred - 17.4% of all resistance cases (4/27) and 3.2% of relapses (2/63). Five patients had KMT2A rearrangements and one had a cryptic t(12;19)(p13;p13)/TCF3::ZNF384 with a normal 46,XX karyotype, found by FISH, RNA-seq and exome sequencing. KMT2A-rearranged patients demonstrated very few additional mutations, with only one of five acquiring pathogenic events in AML relapse, while in the TCF3::ZNF384 case the lineage switch was accompanied by a large set of additional mutations including TP53 p.R89Q and TP53 p.R16H; that patient retained a CD19-positive B population and acquired a CD19-negative myeloid one, scored as acute leukaemia of ambiguous lineage, with leukaemic clone persistence. The authors conclude that lineage switch is not restricted to KMT2A-rearranged cases and can be anticipated in ALL with ZNF384 rearrangement, and recommend careful investigation of all bone-marrow compartments before and after blinatumomab especially in patients with KMT2A rearrangements and ZNF384 fusions.
Lineage Conversion in Pediatric B-Cell Precursor Acute Leukemia under Blinatumomab Therapy. · Semchenkova A, Mikhailova E, Komkov A, Gaskova M, Abasov R, Matveev E, et al., Zerkalenkova E, Popov A · Int J Mol Sci 2022What this evidence cannot show
FULL TEXT READ via Europe PMC fullTextXML for PMC8999738. **n = 1 FOR THE ZNF384 ARM.** The mutation-burden contrast is between one ZNF384 patient and five KMT2A patients and is therefore an observation, not a between-group comparison - no statistic is or could be attached to it. The 17.4% and 3.2% denominators are resistance and relapse cases respectively within one blinatumomab-treated cohort at one centre network, so they are conditional on having been treated with blinatumomab and on having resisted or relapsed; they are NOT a lineage-switch rate for ZNF384-rearranged ALL. Note also that the PubMed abstract for this paper has its gene symbols stripped by italic handling and reads 'Five patients had gene rearrangements; one had translocation' - the symbols KMT2A and TCF3::ZNF384 were recovered from the full text, not guessed.
Evidence that argues against this (1)
- The mutation-burden contrast has an alternative explanation the paper itself supplies: the ZNF384 patient's additional mutations were TP53 variants, and the authors note that TP53 mutation is a recognised feature of relapse generally. So the 'large set of additional mutations' may be a property of that patient's relapse biology rather than of ZNF384-driven lineage switch, and with one case there is no way to separate them. The broader caution is that this graph already records, in ana-two-beats-measured-the-same-lineage-switch-and-got-opposite-answers, that lineage-switch measurements in this project have contradicted each other before - a single case should not become a family-level claim. Semchenkova A et al. 2022
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Myeloid lineage switch following chimeric antigen receptor T-cell therapy has been reported in a patient with TCF3-ZNF384 fusion-positive B-lymphoblastic leukaemia.
Myeloid lineage switch following chimeric antigen receptor T-cell therapy in a patient with TCF3-ZNF384 fusion-positive B-lymphoblastic leukemia. · Oberley MJ, Gaynon PS, Bhojwani D, Pulsipher MA, Gardner RA, Hiemenz MC, Ji J, Han J, O'Gorman MRG, Wayne AS · Pediatr Blood Cancer 2018What this evidence cannot show
**TITLE-LEVEL ONLY.** This citation was recovered from the reference list of the blinatumomab paper (reference 18), where its PMID, DOI, journal, year and full author list are printed; the PMID and DOI were read from that reference entry. **This beat did NOT open the paper itself** and therefore knows nothing about the patient's age, the CAR-T product, the interval to switch, the switched immunophenotype, or whether the fusion was retained - although retention is implicit in calling it a lineage switch. The claim is therefore exactly the title and no more. It is n = 1 and a case report.
Evidence that argues against this (1)
- A title is a weak source and this item is graded moderate for that reason alone. The specific risk of citing from a reference list is the one this project's rule 81 records - a reference list is where wrong identifiers come from - and the mitigation used here was that the PMID and DOI were read together from the same reference entry and are internally consistent with the title. Nobody has published a series large enough to estimate how often lineage switch follows CAR-T in ZNF384-rearranged disease, so the risk is documented as occurring and is not quantified. Hirabayashi S et al. 2021
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The Ponte di Legno Childhood ALL Working Group advise that anti-CD19 CAR-T therapy should be used carefully, if at all, in patients with ZNF384 rearrangements because of the risk of lineage switch to myeloid leukaemia as a cause of relapse, noting that lineage switch after CAR-T has been observed in one patient with TCF3-ZNF384; in the same 218-patient series CD13 was expressed in 38-100% and CD33 in 78-100% of cases across partner subgroups, and two patients with TCF3-ZNF384-positive ALL relapsed more than 10 years after diagnosis.
Clinical characteristics and outcomes of B-ALL with ZNF384 rearrangements: a retrospective analysis by the Ponte di Legno Childhood ALL Working Group. · Hirabayashi S, Butler ER, Ohki K, Kiyokawa N, Bergmann AK, Moricke A, Boer JM, Cave H, Cazzaniga G, Yeoh AEJ, et al., Moorman AV, Manabe A · Leukemia 2021What this evidence cannot show
FULL TEXT READ via Europe PMC fullTextXML for PMC8550960. **THE CAR-T RECOMMENDATION IS EXPERT OPINION IN A DISCUSSION SECTION, NOT A STUDY RESULT** - no patient in this series received CAR-T as part of the reported analysis, and the recommendation rests on one cited case report. The marker ranges are across partner subgroups from Supplementary Table 3, which was not retrieved. MPAL cases were excluded from this series by design, which removes the patients whose lineage ambiguity is most overt.
Evidence that argues against this (1)
- 'Carefully, if at all' is a strong recommendation resting on a single case report, and the counter-consideration is real: CD19-directed therapy is among the most effective salvage options in relapsed B-ALL, and withholding it on the basis of one reported lineage switch could cost more lives than it saves. Nobody has estimated the lineage-switch rate in ZNF384-rearranged patients receiving CD19-directed therapy - the blinatumomab series gives one switch among 182 unselected BCP-ALL patients, which cannot be converted into a subtype-specific rate because the number of ZNF384-rearranged patients in that cohort is not reported. The honest position is that the risk is documented and unquantified, and the recommendation is precautionary rather than evidence-based. Semchenkova A et al. 2022
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
A lineage switch from pro-B acute lymphoid leukaemia to acute myeloid leukaemia has been reported in a case with t(12;17)(p13;q11)/TAF15-ZNF384 rearrangement, predating the immunotherapy era for this subtype.
Lineage switch from pro-B acute lymphoid leukemia to acute myeloid leukemia in a case with t(12;17)(p13;q11)/TAF15-ZNF384 rearrangement. · Grammatico S, Vitale A, La Starza R, Gorello P, Angelosanto N, Negulici AD, De Propris MS, Nanni M, Meloni G, Mecucci C, Foa R · Leuk Lymphoma 2013What this evidence cannot show
**TITLE-LEVEL ONLY, AND GRADED WEAK FOR THAT REASON.** The paper is a closed-access Letter with NO PubMed abstract; Unpaywall reports oa_status 'closed' with no repository copy, Semantic Scholar reports CLOSED, and tandfonline.com returns HTTP 403 to every route tried. This beat therefore knows the patient's age, diagnosis immunophenotype, switched immunophenotype, interval to switch, whether the fusion was retained, and the treatment: NONE OF THEM. The claim above is the title restated. MeSH terms imply an adult female patient but that is metadata, not the paper's words, and no age is quoted here.
Evidence that argues against this (1)
- An unread case report is the weakest kind of evidence and it is included only to establish that ZNF384-associated lineage switch predates immunotherapy - which matters, because if switching only ever followed CD19-directed agents it would be a drug effect rather than a property of the disease. But the title alone cannot even confirm that the fusion was demonstrated in the switched population, which is the criterion that distinguishes a true lineage switch from a second leukaemia. Anyone relying on this item should obtain the text. Semchenkova A et al. 2022
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The mechanistic prior for lineage escape in this subtype is established independently of any therapy: TCF3-ZNF384 and co-occurring alterations were detected in purified blasts, in multiple myeloid- and lymphoid-potent progenitor populations and in a subset of HSCs but not in non-leukaemic T or NK cells, in contrast with Ph-like B-ALL where founding lesions do not reach HSCs; ZNF384 rearrangement was present in 48% of paediatric B/myeloid MPAL cases; and the gene-expression profiles of ZNF384-rearranged B/myeloid MPAL and ZNF384-rearranged B-ALL were indistinguishable.
The genetic basis and cell of origin of mixed phenotype acute leukaemia. · Alexander TB, Gu Z, Iacobucci I, Dickerson K, Choi JK, Xu B, et al., Mullighan CG · Nature 2018What this evidence cannot show
FULL TEXT READ via the PMC author-manuscript rendering of PMC6195459 (Europe PMC fullTextXML returns HTTP 500 for this PMCID). **The HSC fractionation was done on ONE ZNF384-rearranged patient.** The 48% figure does not reconcile with the cohort numbers the paper itself prints - 15 named ZNF384-rearranged cases among 35 B/myeloid MPAL is 42.9%, and the denominator behind 48% is not stated - so it is quoted as published and no fraction should be constructed from it.
Evidence that argues against this (1)
- A stem-cell-resident lesion is a prior for lineage escape, not a demonstration of it, and the same paper contains a result pointing the other way for this subtype specifically: GSEA found ZNF384-rearranged B/myeloid MPAL arrested at a MORE MATURE stage of development than other B/myeloid MPAL, with HSC gene sets negatively enriched in that comparison. A relatively mature arrest and a stem-cell-resident founding lesion are both true here against different comparators, and only the second supports this entry's argument. Alexander TB et al. 2018
- confers resistance to →Blinatumomab in infant KMT2A-rearranged ALL — the largest single outcome gain in thirty years
lineage conversion to CD19-negative myeloid leukaemia; 17.4% of blinatumomab resistance cases across all BCP-ALL, one of six switches was TCF3::ZNF384
- analogous to →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
the only cohort containing both families under one therapy: KMT2A switches were mutation-poor, the single ZNF384 switch carried a large set of additional mutations
extends the CD19-directed lineage-switch risk beyond KMT2A to ZNF384 rearrangement, with an explicit monitoring recommendation naming both
both reported immunotherapy-associated ZNF384 lineage switches were TCF3::ZNF384
the switched ZNF384 patient was scored as acute leukaemia of ambiguous lineage, retaining a CD19-positive B population alongside a new CD19-negative myeloid one
- supports →Flow-cytometric MRD: the fastest readout, and the one a lineage switch walks straight out of
a CD19-gated MRD assay cannot see the escaping population by construction; the authors recommend CD45 and light scatter across all marrow compartments