Oncofusion

start / graph / gene-cd19

Geneunrated confidence 0 of 5: unrated

CD19 - the molecule this disease is treated by, named in 114 entries and written in none of them until now

also called CD19 · B4 · B-lymphocyte antigen CD19 · CD antigen CD19 · P15391

CD19 is a protein on the surface of B cells, and it is the handle every modern immune therapy for this leukaemia grabs. The trouble is that the cancer can let go of the handle - and in this particular leukaemia it does so by becoming a different kind of cancer altogether.

The same thing, for a scientist

CD19 (P15391, 556 aa) is the B-lineage co-receptor targeted by **blinatumomab**, **CD19-directed CAR-T** and the antigen whose loss defines the commonest escape from both. This graph carries `target-cd19-car-t-infant-kmt2a-r`, `drug-blinatumomab`, `mechanism-lineage-switch-kmt2a` and several patient stories in which CD19 is the pivot - and **CD19 was named in 114 entries with no entry of its own** until this one, the top of the queue produced by `scripts/check-named-but-unwritten.py` [E4]. **The escape is not one mechanism but two, and this disease gets the worse one.** Antigen loss in B-ALL generally runs through **CD19 mutation and alternative splicing** - exon 2 skipping produces an N-terminally truncated CD19 that **fails to trigger CART-19 killing** while still supporting the cell [E2]. In **KMT2A-rearranged** disease the commoner route is **lineage switch**: the blasts stop being B cells. *"Most commonly, LS has been associated with KMT2A-rearranged leukemias ... including in up to 70% of infant B-ALL in which lineage plasticity is well established"* [E1]. **So CD19-directed therapy in this disease is selecting against a lineage, not against an epitope**, and the two escapes have different consequences: a truncated CD19 is still a B-ALL and can be re-targeted, while a switched AML is a different disease with dismal outcomes.

Go deeper
What the escape looks like, and why the distinction is therapeutic

Project EVOLVE collected 75 lineage switches after antigen-targeted therapy - against CD19, CD22, CD38 and CD7 - and the split is 53 (70.7%) B-ALL to AML, 17 (22.7%) to MPAL or acute leukaemia of ambiguous lineage, and 5 rarer presentations 1. The paper's own framing puts KMT2A rearrangement at the centre: it is "found in both AML and ALL" and KMT2A-rearranged B-ALL has a more stem cell-like phenotype, with AML emerging even after B-ALL-directed chemotherapy alone.

The infant case that anchors it in this graph is older and smaller and points the same way: a t(4;11) infant refractory to chemotherapy, treated with blinatumomab, clearing peripheral lymphoblasts and then switching to CD19-negative monoblastic AML 3.

Why the two escapes are not interchangeable. A CD19 splice variant or a point mutation leaves a B-ALL that has lost an epitope - CD22-directed therapy, or a CAR against a different CD19 domain, remains on the table. A lineage switch leaves a myeloid leukaemia, and every B-lineage agent in the protocol is now inert. This graph's MRD beat found the sharper version of the same problem: a B-lineage flow panel scores a switched blast as remission.

Cell type, cell state, localisation, epigenetics
  • Cell type. B lineage from pro-B onward, lost on plasma-cell differentiation. In this disease the relevant fact is narrower: KMT2A::AFF1 infant B-ALL is characteristically CD10-negative pro-B, and CD19 is the marker that survives that immaturity - which is exactly why it became the target.
  • Cell state. The escape is a state transition, not a mutation, in the KMT2A-rearranged case. cell-state-under-menin-inhibition and mechanism-lineage-switch-kmt2a carry the surrounding argument; what CD19 adds is that the selection pressure is iatrogenic.
  • Subcellular. Plasma membrane, in complex with CD21, CD81 and CD225 as the B-cell receptor co-receptor; the cytoplasmic tail carries the signalling tyrosines 5.
  • Subnuclear. Not applicable - and worth saying, because this graph's other gene entries are nuclear proteins and a reader scanning for a subnuclear line should find the reason it is absent rather than an omission.
  • Epigenetics. The link to this graph's subject runs the other way from usual: CD19 expression is a readout of the B-lineage programme that the fusion holds in place or releases, not a target of the fusion's chromatin machinery.
Facts and figures
uniprotP15391
gene symbolCD19
length AA556
subnuclear localizationnot applicable - CD19 is a single-pass plasma membrane protein and no subnuclear pool has been reported
shown for fusionNO - neither localisation study states the KMT2A status of its cases, and neither analysed a KMT2A-rearranged cohort. The surface compartment IS routinely read in KMT2A-rearranged infant B-ALL by diagnostic flow cytometry, but the trafficking mechanism has not been demonstrated in that subtype.
localization
{
  "subcellular": "plasma membrane, type I single-pass (P15391, 556 aa). The therapeutically decisive statement is not that CD19 is a membrane protein but that it must REACH the outer leaflet: both characterised escape routes after CD19-directed therapy leave the gene intact and the protein intracellular - a post-endoplasmic-reticulum export block after blinatumomab, and endoplasmic-reticulum retention of a misfolded exon-2-altered protein after CAR-T.",
  "subnuclear": "not applicable - CD19 is not a nuclear protein and no nuclear pool has been reported in any source read for this beat",
  "how_shown": "molecular workup of a CD19-negative blinatumomab relapse localising the block to the post-endoplasmic-reticulum compartment (PMID 27784674); biochemical demonstration of endoplasmic-reticulum retention of post-CAR-T CD19 variants (PMID 30104252). Routine diagnosis reads this compartment by unpermeabilised flow cytometry.",
  "systems": "B-lineage acute lymphoblastic leukaemia patients and cell models under CD19-directed immunotherapy; fusion status not stated in either source",
  "node": "localization-plasma-membrane-leukaemic-surface-antigens"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.016,
  "chronos_mean_other_leukaemia": -0.1894,
  "difference": 0.1733,
  "pan_essentiality_all_lines": -0.0761,
  "selectivity_rank": 17542,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
How this could bite us back
  • CD19-directed therapy in KMT2A-rearranged disease selects for lineage switch, and a switched blast is a different leukaemia with dismal outcomes - not a relapse of the one being treated. Any proposal to intensify or extend CD19-directed therapy in this genotype has to state what it will do about that.
  • The escape is invisible to the assay that defines response. A B-lineage flow panel scores a switched myeloid blast as remission, and this graph's MRD beat found that no menin-inhibitor trial has reported paired phenotype-based and lineage-independent molecular MRD. The same gap applies to CD19-directed therapy.
  • Targeting a second antigen is not automatically an answer. Project EVOLVE collected switches after CD19, CD22, CD38 AND CD7 targeting, so the phenomenon is not specific to the CD19 epitope and a dual-antigen strategy may select the same way.
  • Everything here is about the leukaemia and nothing about the patient's B cells. CD19-directed therapy ablates normal B lineage as well, and the population is infants whose humoral immunity is still developing - a toxicity this entry does not characterise and that any therapeutic argument must.
  • Localisation resistance is invisible to the commonest resistance assays. A CD19-negative relapse driven by a post-endoplasmic-reticulum export block still has CD19 mRNA and still has CD19 protein in the lysate, so an RNA-seq or western-blot readout of a resistance model will report the target as present. Only an unpermeabilised surface stain sees the escape (PMID 27784674, PMID 30104252).
Still open
  1. What fraction of CD19 escapes in KMT2A-rearranged infants is lineage switch versus epitope loss? Project EVOLVE has the cases and does not resolve infants with t(4;11) separately.
  2. Is the switched clone pre-existing or induced? A case report cannot distinguish them, and the answer changes whether the right intervention is earlier myeloid-directed therapy or a different antigen.
  3. Does CD19 exon 2 skipping occur in KMT2A-rearranged B-ALL at all? The splicing route is characterised in general B-ALL and this graph has a splicing-landscape hypothesis with no CD19 arm.
  4. Write `gene-cd34`, which tops the same queue at 127 entries and defines the cell of origin in every experiment in this graph.
Every source behind this entry (7)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.journal-article

    Across 75 international cases of lineage switch after antigen-targeted immunotherapy (CD19, CD22, CD38, CD7), 53 (70.7%) were B-ALL transforming to AML, 17 (22.7%) to MPAL or acute leukaemia of ambiguous lineage, and 5 were rarer presentations. Lineage switch is most commonly associated with KMT2A-rearranged leukaemias, a cytogenetic alteration found in both AML and ALL and present in up to 70% of infant B-ALL, in which lineage plasticity is well established; KMT2A-rearranged B-ALL has a more stem-cell-like phenotype and AML has emerged after B-ALL-directed chemotherapy alone.

    PMID 40193715
    What this evidence cannot show

    Read in full on 2026-09-20 (PMC12333221, 53,342 characters after stripping) and the KMT2A passage quoted verbatim. **Established as what this series reports.** It is a **retrospective international collection of reported cases**, so it has no denominator: it cannot give an incidence of lineage switch, only a description of the cases that were collected and shared. The KMT2A-rearranged framing is the authors' introduction citing prior work, not a result of this series.

    Evidence that argues against this (1)
    • A collection of reported lineage switches will over-represent the genotype people already expect to switch, because a switch in a KMT2A-rearranged patient is more likely to be recognised, worked up and reported than the same event in a genotype nobody associates with plasticity. The association is real in this graph on other grounds, but this series cannot establish its strength. Rayes A et al. 2016
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.journal-article

    Relapse with CD19 epitope loss occurs in 10-20% of paediatric responders to CD19-directed CAR-T. Mechanisms include hemizygous deletions spanning the CD19 locus, de novo frameshift and missense mutations in exon 2, and alternatively spliced CD19 mRNA including a species lacking exon 2; SRSF3 was identified as a splicing factor involved in exon 2 retention and its levels were lower in relapsed B-ALL. Exon 2 skipping bypasses exon 2 mutations and allows expression of an N-terminally truncated CD19 that fails to trigger killing by CART-19.

    PMID 26516065
    What this evidence cannot show

    Read from the abstract on 2026-09-20; **full text not read**. Established as a result within the reported study. **This is B-ALL generally, not KMT2A-rearranged specifically**, and the 10-20% figure is for paediatric responders to one CAR construct in one era. Nothing here establishes the rate of splice-mediated escape in infant KMT2A::AFF1, where the competing route is lineage switch.

    Evidence that argues against this (1)
    • If lineage switch dominates in KMT2A-rearranged disease [E1], then the splice-and-mutation route this paper characterises may be the MINOR escape here even though it is the better-characterised one - and a resistance panel built from the general B-ALL literature would look for the wrong thing in this genotype. Silbert SK et al. 2025
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.journal-article

    An infant with t(4;11) ALL refractory to cytotoxic chemotherapy was treated with blinatumomab, cleared peripheral lymphoblasts rapidly, and on bone marrow evaluation was found to have switched to CD19-negative monoblastic AML.

    Lineage Switch in MLL-Rearranged Infant Leukemia Following CD19-Directed Therapy · Rayes A, ..., O'Brien MM · Pediatric Blood & Cancer 2016
    PMID 26914337
    What this evidence cannot show

    Read from the abstract on 2026-09-20. **A single case report**, which establishes that the event occurs in the index fusion in an infant and establishes nothing about its frequency. It is carried because this graph's index population is infants with t(4;11) and the larger series [E1] contains such patients without resolving them separately.

    Evidence that argues against this (1)
    • A case report selected for being publishable cannot distinguish a switch caused by the therapy from a pre-existing myeloid subclone unmasked by it - a distinction this graph treats as open elsewhere, and one the case report format is structurally unable to settle. Silbert SK et al. 2025
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis

    CD19 was named in the text of 114 entries of this graph with no entry of its own, the second-highest of 171 HGNC symbols in that state and behind only CD34 at 127.

    What this evidence cannot show

    Measured on disk 2026-09-20 while two agents were writing, so it is a snapshot. **The count is of entries mentioning the symbol, not of mentions**, and frequency is not importance - the instrument's own entry says so. CD19 is written here because of what it is in this disease, not because it topped a list.

    Evidence that argues against this (1)
    • The same instrument ranks REST at 195 and EFS at 123, both of which are ordinary words rather than the genes HGNC names. A rank from it is a reason to look and never a reason to write. Oncofusion knowledge graph (internal) 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    CD19 is UniProt P15391 (CD19_HUMAN), B-lymphocyte antigen CD19, 556 amino acids, with a single reviewed human record and synonyms including B4 and CD antigen CD19.

    What this evidence cannot show

    Fetched 2026-09-20 with `gene_exact` and `reviewed:true`, `gene_primary` and `length` requested and the primary name read back, per rules 62 and 65. Exactly one reviewed record was returned, so the paralogue-collision hazard that affects HDAC7/HDAC9 and CSNK2A1/CSNK2A2 does not arise here.

    Evidence that argues against this (1)
    • A canonical length is not the protein the therapy sees. The escape variant in [E2] is an N-terminally truncated CD19 produced by exon 2 skipping, so the 556-residue canonical form is the wrong reference for any argument about epitope retention. Sotillo E et al. 2015
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In B-lineage ALL relapsing on the CD19/CD3 BiTE blinatumomab, CD19 escape can be a localisation failure rather than an expression failure: molecular workup of one early relapse identified a disrupted CD19 membrane export in the post-endoplasmic-reticulum compartment, with the rest of the immunophenotype unchanged from diagnosis.

    Resistance to anti-CD19/CD3 BiTE in acute lymphoblastic leukemia may be mediated by disrupted CD19 membrane trafficking. · Braig F, Brandt A, Goebeler M, Tony HP, Kurze AK, Nollau P, Bumm T, Böttcher S, Bargou RC, Binder M · Blood 2017
    PMID 27784674DOI
    What this evidence cannot show

    Four relapses reported, ONE worked up mechanistically - a single-case molecular result, and the paper's own title hedges with 'may be mediated by'. The KMT2A status of the cases is not stated in the abstract and was not resolvable for this beat, so per brief rule 73 this is a B-ALL result and not a KMT2A-rearranged one. It is also not a microscopy experiment: the compartment assignment is biochemical.

    Evidence that argues against this (1)
    • The dominant published mechanism of CD19 escape is genetic and splicing-level rather than trafficking-level: acquired mutations in CD19 together with exon-2 skipping are selected under CART-19 pressure and produce a truncated protein. The two are not exclusive - the exon-2 product is itself retained - but a claim that CD19 escape is a trafficking problem overstates the trafficking paper, which is n=1. Sotillo E et al. 2015
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CD19 alterations emerging after CD19-directed immunotherapy cause retention of the misfolded CD19 protein in the endoplasmic reticulum, so the escape cell transcribes CD19 and scores CD19-negative on an unpermeabilised surface stain.

    CD19 Alterations Emerging after CD19-Directed Immunotherapy Cause Retention of the Misfolded Protein in the Endoplasmic Reticulum. · Bagashev A, Sotillo E, Tang CH, Black KL, Perazzelli J, Seeholzer SH, Argon Y, Barrett DM, Grupp SA, Hu CC, Thomas-Tikhonenko A · Mol Cell Biol 2018
    PMID 30104252DOI
    What this evidence cannot show

    This record carries a published correction (PMID 35976030, Mol Cell Biol 2022) whose content was NOT read for this beat; check it before quoting any number from this paper. System is CD19-directed immunotherapy in B-ALL, not specifically infant KMT2A-rearranged disease.

    Evidence that argues against this (1)
    • Lineage switch is the competing explanation that is specifically enriched in KMT2A-rearranged disease, and it is not a localisation event at all - the B-cell programme including CD19 is shut down wholesale. The trafficking series explicitly excludes lineage switch in its own cases on the grounds that the rest of the immunophenotype was preserved, which is exactly the control a lineage-switch case fails. Braig F et al. 2017
Connected to