The leukaemia that changed lineage in eleven days and changed back when the drug stopped: KMT2A::AFF4 infant ALL on blinatumomab
also called blinatumomab lineage switch reversion case · KMT2A-AFF4 ins(5;11) infant ALL case
A girl diagnosed with B-cell leukaemia at five months old relapsed after a transplant and was given blinatumomab, an antibody that directs the immune system against B cells. Within eleven days her leukaemia had stopped being a B-cell leukaemia and become a myeloid one - and when the drug was stopped it turned back into a B-cell leukaemia within days; she later developed solid deposits of leukaemia outside the marrow and died.
The same thing, for a scientist
A girl diagnosed at 5 months of age with pro-B ALL carrying a KMT2A::AFF4 fusion generated by an atypical insertion ins(5;11), with a blast phenotype of surface CD19 bright, NG2 bright and CD22 positive, relapsed 16 months after allogeneic transplant with 18,000 blasts/uL. After cytoreduction she entered a blinatumomab expanded-access programme. By day 11 the marrow showed almost complete disappearance of lymphoblasts (5%) and the emergence of monocytic cells (80%), with a flow phenotype of surface CD19 low, CD33+, CD34-, CD14++, CD15++, CD11b++, CD64+ consistent with an M5 AML - although the NG2-directed antibody 7.1 remained at least partially positive, marking the clone as the same one. Blinatumomab was stopped without substituting chemotherapy; the circulating blasts disappeared within days, and a repeat marrow nine days later showed reversion to approximately 40% lymphoblasts with only rare monocytic cells. Subsequent chemotherapy and a stem cell boost left 8% residual blasts. She then developed aggressive extramedullary disease with chloromas and died despite palliative intervention and reintroduction of blinatumomab. [E1]
How it began, and what was found
The published record for this child begins at diagnosis, not at first symptom: the report is a short letter and does not describe what the family noticed. It records a diagnosis of pro-B ALL at 5 months of age, with a KMT2A::AFF4 fusion arising from an atypical insertion, ins(5;11) rather than a reciprocal translocation, and a blast phenotype of surface CD19 bright, NG2 bright and CD22 positive 1.
Two details are worth pulling out.
AFF4, not AFF1. AFF4 is a member of the same AF4/FMR2 family and a core component of the super elongation complex; it is a rare fusion partner in infant ALL compared with AFF1 at t(4;11). The graph carries gene-aff4 as a component of the elongation machinery; here it is the 3' partner of the fusion itself.
An insertion, not a translocation. ins(5;11) will not necessarily look like a translocation on a standard karyotype, which is one of the concrete reasons concept-fish-split-signal-kmt2a and concept-molecular-confirmation-fusion-transcript exist as separate steps from concept-cytogenetics-karyotype-kmt2a.
NG2 bright. NG2/CSPG4 is the proteoglycan repeatedly associated with KMT2A-rearranged blasts and with extramedullary infiltration. It is the marker that made this case interpretable, because it was still detectable after the lineage had changed.
What was done and what happened
She was treated with chemotherapy and an allogeneic transplant, and relapsed 16 months after transplant with 18,000 blasts/uL. After cytoreduction she entered a blinatumomab expanded-access programme 1.
Day 11. The marrow showed "almost complete disappearance of the lymphoblastic cells (5%), but the occurrence of monocytic cells (80%)". Flow cytometry gave surface CD19 low, CD33+, CD34-, CD14++, CD15++, CD11b++, CD64+ - an M5 monocytic AML phenotype. The antibody 7.1, which detects NG2, was at least partially positive, which is what identifies the monocytic population as the same clone rather than a second leukaemia.
Then the drug was stopped, and nothing else was given. Within days of ending blinatumomab the blasts disappeared from the peripheral blood. A repeat marrow nine days later showed reversion: approximately 40% lymphoblastic blasts and only rare monocytic cells.
That is the finding. A single clone moved from a lymphoid to a monocytic phenotype under CD19-directed immune pressure in eleven days, and moved back when the pressure was removed, with no cytotoxic therapy given in between.
Outcome. Subsequent chemotherapy and a stem cell boost left 8% residual blasts. She then developed aggressive extramedullary disease with chloromas and died, despite palliative intervention and reintroduction of blinatumomab 1.
Why this case is in the graph1 deeper
risk-lineage-switch-cd19-directed-therapy states the risk. This case is the cleanest published demonstration of its mechanism, and the reason is the reversion.
If a lineage switch were selection - a pre-existing myeloid subclone outgrowing a lymphoid one because the lymphoid one is being killed - then removing the pressure would not restore the lymphoid phenotype in nine days. Outgrowth is not reversible on that timescale. Reversion at nine days is evidence of reprogramming: the same cells changing what they express, not different cells winning. The authors' own summary describes it as leukaemia that "rapidly adapts to immunological stimuli illustrating the high plasticity of this phenotype".
Three further things a category entry cannot give you:
- NG2 was the thread. Without a marker that survives the switch, this would have been recorded as a therapy-related second leukaemia. Clonality evidence is what turns an anecdote into a mechanism.
- The switch is not the end of the story, and reversion is not rescue. The phenotype came back and the child still died, of extramedullary disease.
- Escape was phenotypic, not genetic. No resistance mutation was required. That is a different, and harder, problem than the MEN1 binding-site mutations the graph catalogues for menin inhibitors.
What this implies for any surface-antigen-directed therapy in this disease
CD19, CD22 and CD33 are all expression states, and this case shows the state is negotiable within days. A menin inhibitor is aimed at something the fusion cannot switch off without ceasing to be oncogenic, which is an argument in its favour - but the same plasticity that lets a clone drop CD19 is what makes differentiation-based readouts hard to interpret in this disease, and it applies to menin inhibitors too, because their proposed mechanism is differentiation.
| age at dx | 5 months |
|---|---|
| sex as reported | female |
| fusion | gene-aff4 |
| fusion description | KMT2A::AFF4 arising from an atypical insertion ins(5;11), not a reciprocal translocation; no dedicated fusion node exists in this graph for KMT2A::AFF4, so the edge is made to gene-aff4 |
| cns status | not stated in the report |
| course | chemotherapy and allogeneic transplant; relapse 16 months after transplant with 18,000 blasts/uL; cytoreduction then blinatumomab under an expanded-access programme; day-11 marrow 5% lymphoblasts and 80% monocytic cells with M5 AML phenotype; blinatumomab stopped with no substituted chemotherapy; peripheral blasts cleared within days and a marrow 9 days later showed approximately 40% lymphoblasts with only rare monocytic cells; subsequent chemotherapy and stem cell boost left 8% residual blasts |
| outcome | died, after developing aggressive extramedullary disease with chloromas, despite palliative intervention and reintroduction of blinatumomab; the report does not state the interval to death |
| key observation | the lineage switch reversed within nine days of stopping the drug, with no cytotoxic therapy in between, which is consistent with phenotypic reprogramming of one clone rather than outgrowth of a pre-existing myeloid subclone |
| source type | case report |
| consent note | Published peer-reviewed case report in a haematology journal. This entry records only what the publication discloses; the treating institutions are deliberately omitted from the body text. |
| privacy review | No detail added, inferred or reconstructed beyond the publication. The death is recorded as reported. Because the letter discloses very little about the child, this entry is correspondingly sparse rather than filled in from typical-case knowledge. |
presentation
[ "the report does not describe the presenting symptoms at diagnosis; it is a short letter and begins at the diagnosis of pro-B ALL" ]
immunophenotype at diagnosis
[ "surface CD19 bright", "NG2 bright", "CD22 positive" ]
immunophenotype after switch
[ "surface CD19 low", "CD33+", "CD34-", "CD14++", "CD15++", "CD11b++", "CD64+", "NG2 (antibody 7.1) at least partially positive" ]
- Any therapy directed at a surface antigen in KMT2A-rearranged leukaemia is aimed at a state the cell can leave within days and re-enter within days. Escape here requires no mutation, so nothing in a resistance-mutation screen will predict it.
- Because the phenotype is reversible, an apparent response to a lineage-directed agent and an apparent switch are both unstable readouts. A trial endpoint that depends on immunophenotype in this disease is measuring something the disease can change on purpose.
- Menin inhibitors are proposed to work by releasing the differentiation block. This case shows the differentiation programme in KMT2A-rearranged blasts is unusually labile in both directions, which cuts both ways: it may make differentiation therapy effective, and it may make differentiation syndrome and phenotypic drift harder to control.
- The child died of chloromas - extramedullary myeloid deposits - after the marrow had been repeatedly reduced. A drug evaluated on marrow blast percentage would have looked better than the patient was.
- How fast is the switch, really? This case caught it at day 11 and the reversion at day 9 after stopping, but nobody has sampled daily. Serial single-cell profiling during the first two weeks of CD19-directed therapy in a KMT2A-rearranged patient would give the kinetics, and the kinetics distinguish reprogramming from selection more decisively than any endpoint marker.
- Is NG2/CSPG4 a lineage-independent clonal marker in KMT2A-rearranged leukaemia, and should it therefore be a standard part of the MRD panel in any patient receiving lineage-directed immunotherapy?
- This clone reverted to lymphoid when the pressure was removed, and the child then died of extramedullary myeloid disease. Does the switched state have a different tissue tropism from the unswitched one, and does CD19-directed therapy therefore select for extramedullary relapse specifically?
- KMT2A::AFF4 versus KMT2A::AFF1: both fuse KMT2A to an AF4-family super-elongation-complex subunit. Is the disease biology, including this plasticity, the same, or does the partner matter? There are too few AFF4 cases to answer and the question is open.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A girl diagnosed at 5 months of age with pro-B ALL carrying a KMT2A::AFF4 fusion from an atypical ins(5;11), with blasts bright for surface CD19 and NG2 and positive for CD22, relapsed 16 months after allogeneic transplant with 18,000 blasts/uL and received blinatumomab; by day 11 the marrow showed almost complete disappearance of lymphoblasts (5%) with 80% monocytic cells of M5 AML phenotype (surface CD19 low, CD33+, CD34-, CD14++, CD15++, CD11b++, CD64+) that remained at least partially positive for the NG2-directed antibody 7.1; on stopping blinatumomab without substituting chemotherapy the blasts disappeared from the periphery within days and a marrow nine days later showed reversion to approximately 40% lymphoblasts with only rare monocytic cells; subsequent chemotherapy and a stem cell boost left 8% residual blasts, and the patient developed aggressive extramedullary disease with chloromas and died despite palliative intervention and reintroduction of blinatumomab.
Spontaneous reversion of a lineage switch following an initial blinatumomab-induced ALL-to-AML switch in KMT2A-rearranged infant ALL · Wolfl M, Rasche M, Eyrich M, Schmid R, Reinhardt D, Schlegel PG · Blood Advances 2018What this evidence cannot show
Full text read via the rendered PubMed Central page (PMC6020817); the PubMed abstract for this item is only a one-line key point, and the PMC plain-text extract returned empty, so the case details here come from reading the rendered article. It is a short letter, so several things a case report would normally give are absent: there is no description of the presenting symptoms at 5 months of age, no blood count at diagnosis, no CNS status, no conditioning regimen and no stated interval from the chloromas to death. A single case: it demonstrates that reversible lineage switching happens, and establishes no frequency; publication bias is total here, since spontaneous reversion is the only reason this case was published. The clonal identity of the lymphoid and monocytic populations rests on partial positivity of a single NG2-directed antibody rather than on sequencing of the breakpoint or of immunoglobulin rearrangements, so the reprogramming interpretation is strong but not proven in this patient.
Evidence that argues against this (2)
- Against concluding that every lineage switch is reprogramming rather than selection: a case with the same phenomenon studied by clonal IGH/IGL sequencing found three dominant sequences shared across the original B-ALL, the switched AML and the post-transplant relapse - which is consistent with reprogramming, but that study also found additional sequences at relapse that were present at low abundance in the diagnostic sample, i.e. evidence that subclonal selection operates alongside it. source 2021
- Against reading blinatumomab as generally unsafe in this genotype: the same drug, given with a CD33-directed conjugate to co-target a pre-existing myeloid clone in a KMT2A::AFF1 infant, produced an MRD-negative remission that supported transplant, with the child alive in second remission at 36 months of age. source 2021
the index-population case
KMT2A::AFF4 fusion from an atypical ins(5;11); no KMT2A::AFF4 fusion node exists in this graph
pro-B ALL diagnosed at 5 months of age
switch to M5 AML phenotype by day 11 of blinatumomab, reverting within 9 days of stopping the drug with no cytotoxic therapy given
- treated by →Blinatumomab in infant KMT2A-rearranged ALL — the largest single outcome gain in thirty years
given under an expanded-access programme for post-transplant relapse; precipitated and, on withdrawal, released the lineage switch
the immunophenotype changed lineage and changed back in one patient, with NG2 the only marker that persisted through both
an insertion ins(5;11) rather than a reciprocal translocation, which a standard karyotype need not reveal
relapse 16 months after allogeneic transplant, then a stem cell boost that left 8% residual blasts
two infants given blinatumomab for KMT2A-rearranged leukaemia: one switched lineage and died, one had its myeloid clone co-targeted and survived into a second remission
two lineage switches in KMT2A-rearranged disease with opposite provocations and opposite outcomes: spontaneous under chemotherapy with long survival, versus drug-induced under blinatumomab with death
reversion in nine days with no intervening therapy excludes subclonal outgrowth in that case