Oncofusion

start / graph / gene-cd34

Genestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q45-immunophenotype

CD34 - named in 132 entries of this graph and written in none, and it does not mean what the graph has been using it to mean

also called CD34 · Hematopoietic progenitor cell antigen CD34 · CD antigen CD34 · P28906

CD34 is the sticker scientists use to pull haematopoietic stem and progenitor cells out of blood or marrow, and almost every experiment in this graph starts by sorting CD34-positive cells. In the leukaemia this graph is about, CD34 does not reliably mark the cells that start or sustain the disease - some leukaemia-initiating cells are CD34-negative, and some CD34-positive cells in the same marrow are perfectly normal stem cells.

The same thing, for a scientist

CD34 (UniProt **P28906**, CD34_HUMAN, **385 aa**, single reviewed human record, no gene synonyms) is a sialomucin on haematopoietic stem and progenitor cells and on vascular endothelium. It tops `scripts/check-named-but-unwritten.py` at **132 entries naming it with no entry of its own** [E6], which is unsurprising: it is the sorting gate for `cell-type-fetal-liver-hsc`, `exp-human-cd34-kmt2a-breakage-assay`, `model-murine-af4-in-human-cd34-pro-b-all` and essentially every cord-blood transduction model in the graph. **The load-bearing fact is that CD34 fractionates the wrong thing in this disease.** Aoki 2015 xenografted sorted fractions from primary infant MLL-rearranged leukaemias: in **MLL-AF4 patients both CD34+CD38+CD19+ and CD34-CD19+ cells initiated leukaemia**; in MLL-AF9 patients **CD34-CD19+ cells were the LICs**; in MLL-ENL patients either could be, depending on the case's CD34 pattern. Meanwhile, in those same patients, **CD34+CD38-CD19-CD33- cells were enriched for normal HSCs** with long-term multilineage repopulation and polyclonal IGH [E1]. So in the index fusion, sorting CD34-positive does not purify the leukaemia-initiating compartment, and the most stem-like CD34 gate is where the *normal* cells are. **And CD34 positivity does not identify this disease at diagnosis either.** In adult B-precursor ALL, NG2 positivity - the KMT2A surrogate - is significantly associated with a **CD34-negative** phenotype [E2]; in a 1,382-case B-ALL series, KMT2A-rearranged cases had significantly **lower** CD34 mean fluorescence intensity than KMT2A-germline cases [E3]. A 2026 Blood image report describes **bimodal CD34 expression** in adult KMT2A-rearranged pro-B ALL [E4]. The graph's own `concept-flow-cytometry-immunophenotype-kmt2a-r` says these cases are "usually CD34-positive", and that is the paediatric convention rather than a general truth. **Where CD34 does carry information is as an escape reservoir.** Bueno 2022 found **CD34+CD19-CD22+ progenitors in the marrow of about 70% of B-ALL patients**, at threefold higher pre-treatment frequency in those who later relapsed after CD19-directed immunotherapy, carrying the diagnostic genetic abnormality and initiating leukaemia in xenografts [E5]. That is a CD34-gated population defined by what it *lacks*.

Go deeper
What CD34 positivity does and does not mean in infant KMT2A-rearranged disease

Does mean:

  • The cell sits in the haematopoietic stem/progenitor compartment by surface phenotype, which is why CD34 selection is the right starting material for a transformation experiment (exp-human-cd34-kmt2a-breakage-assay, model-murine-af4-in-human-cd34-pro-b-all).
  • In paediatric pro-B KMT2A-r ALL it is commonly co-expressed with CD19 and TdT, and it contributes to the EGIL BI pattern.
  • In a CD19-directed-therapy context, a CD34+CD19- fraction is where phenotypic escape is pre-loaded 5.

Does not mean:

  • Not the leukaemia-initiating cell. In MLL-AF4 both CD34+ and CD34- fractions initiate; in MLL-AF9 the CD34- fraction does 1. A CD34-purified graft in this disease is not a purified LIC graft.
  • Not a purity gate against normal cells. The CD34+CD38-CD19-CD33- fraction of a patient with this leukaemia contains normal HSCs that repopulate with polyclonal IGH 1. Any experiment reading a CD34+ fraction as "the leukaemia" is reading a mixture.
  • Not a diagnostic feature of the genotype. KMT2A-r associates with lower CD34 MFI and, in adults, with CD34-negativity 234.
  • Not a prognostic constant. In a 105-child B-ALL series CD34-positivity (>=10%) was associated with lower MRD and relapse, but that series' CD34-negative group was enriched for MLL-AF4 and TCF3-PBX1, so the CD34 effect and the genotype effect are confounded 7.
Cell type, cell state, localisation, epigenetics
  • Cell type. Haematopoietic stem and progenitor cells; also vascular endothelium and some mesenchymal populations, which is why CD34 is not a haematopoiesis-specific marker and why a CD34-directed agent would not be a haematopoiesis-restricted agent.
  • Cell state. CD34 expression is a state, not an identity - it is downregulated on differentiation and is dynamic under therapy. This is the property that makes it a poor MRD anchor and a good enrichment gate.
  • Localisation. Type I transmembrane sialomucin at the plasma membrane; 385 residues 8. No subnuclear address.
  • Epigenetics. No entry in this graph, and none found here, establishes CD34 as a direct KMT2A-fusion target gene - unlike gene-cspg4 and gene-prom1, both of which are. The correct statement is that CD34 tracks the differentiation stage the fusion arrests cells at, not that the fusion drives the CD34 locus.
Facts and figures
uniprotP28906
gene symbolCD34
length AA385
cd numberCD34
accession resolutiongene_exact:CD34 AND organism_id:9606 AND reviewed:true, fields include gene_primary and length; one row; primary name read back as CD34; length 385 asserted for verify-accessions.py.
subnuclear localizationNOT APPLICABLE. CD34 is a single-pass type I membrane protein (UniProt P28906); there is no nuclear pool and therefore no subnuclear compartment to record. Distinguished deliberately from the unpublished case - for SETD1A or EZH1 the subnuclear question is answerable and unasked, whereas here it is ill-posed.
shown for fusionYES for the compartment, NO for anything finer, and the two must not be confused. CD34 is detected at the surface of KMT2A-rearranged blasts constantly - that is what the immunophenotyping literature in this entry is - and this entry already records bimodal CD34 expression in adult KMT2A-rearranged pro-B ALL [E4] and CD34-positive and CD34-negative leukaemia-initiating cells in infant MLL-rearranged ALL [E1]. What has never been done is a cell-biological localisation study of CD34 protein in a KMT2A-rearranged cell: measured 2026-09-21 with NCBI esearch, CD34[tiab] AND (MLL[tiab] OR KMT2A[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab] OR mixed lineage leukemia[tiab]) AND (localiz*[tiab] OR localis*[tiab] OR subcellular[tiab] OR nuclear body[tiab] OR nuclear bodies[tiab] OR speckle*[tiab] OR condensat*[tiab] OR nucleolar[tiab] OR immunofluorescence[tiab] OR compartment*[tiab]) returns 10 records, all title-checked off-topic for CD34 protein localisation - they are UBTF duplications, microenvironment lineage fate, HOXA10-AS, IL-3/TPO stromal scaffolds, an unrelated Chinese-language angiogenesis paper, ZNF521, flavonoid-induced MLL translocations, NPM-mutant immunophenotype, molecular relapse detection and leukapheresis. Operands: gene clause 41,855, gene AND fusion 178. Positive control through the identical shape with (MLLT1[tiab] OR ENL[tiab]) returns 23. Zero in titles and abstracts; supporting figures not searched (brief rule 80).
lic fractionation mll
{
  "MLL-AF4": "CD34+CD38+CD19+ and CD34-CD19+ both initiate",
  "MLL-AF9": "CD34-CD19+ initiate",
  "MLL-ENL": "CD34+ or CD34- depending on the case's CD34 pattern",
  "normal_HSC_fraction_in_same_patients": "CD34+CD38-CD19-CD33-",
  "pmid": "25538041",
  "title": "Identification of CD34+ and CD34- leukemia-initiating cells in MLL-rearranged human acute lymphoblastic leukemia.",
  "authors": "Aoki Y, Watanabe T, Saito Y, et al.",
  "venue": "Blood",
  "year": 2015,
  "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
named but unwritten rank
{
  "entries_naming": 132,
  "rank": 1,
  "queue_size": 163,
  "measured": "2026-09-20"
}
localization
{
  "subcellular": "PLASMA MEMBRANE, and that is the whole answer. UniProt P28906 gives one location - Membrane, single-pass type I membrane protein - with no evidence code and no note. CD34 is a type I transmembrane glycoprotein whose extracellular domain is heavily glycosylated and whose short intracellular tail is phosphorylated. It is the only entry in this tranche that is not a nuclear protein, and the only one a therapeutic antibody could reach.",
  "subnuclear": "NOT APPLICABLE - and this is a genuine not-applicable, not an unpublished gap. A single-pass type I membrane protein with an extracellular glycosylated domain and a cytoplasmic tail has no nuclear pool to sub-localise. Nothing should be inferred from the absence, and no query was run for a subnuclear CD34 pool because the question is ill-posed.",
  "how_shown": "UniProt P28906 SUBCELLULAR LOCATION comment; topology and domain structure restated with mass-spectrometric mapping of intracellular-domain phosphorylation (PMID 21499536); the operational evidence for the compartment is the entire flow-cytometry and immunomagnetic-selection literature, which detects CD34 on intact unpermeabilised cells.",
  "systems": "mouse CD34 in HEK293F for the phosphorylation mapping; the surface-detection evidence is human and clinical but is antibody-based rather than cell-biological",
  "node": "localization-plasma-membrane-leukaemic-surface-antigens"
}
cell type
{
  "expressed_in": "Haematopoietic stem and progenitor cells, endothelial cells and mast cells, and in mouse also hair-follicle epidermal stem cells and human outer root sheath cells [E_q66_loc1] - CD34 is NOT haematopoiesis-specific, which is the first thing a graph that treats it as a stemness marker should record. In this disease: 82.9 percent of 105 children with newly diagnosed B-ALL had bone marrow CD34-positive cells at or above 10 percent [E7], and adult KMT2A-rearranged pro-B ALL shows BIMODAL CD34 expression [E4].",
  "required_in": "No cell requires CD34 for viability. What it does instead is bias lineage output: silencing it in human HSCs enhances granulocyte and megakaryocyte differentiation and reduces erythroid maturation, and overexpression does the reverse [E_q66_cell1]. That is a function, and it is not the function the marker is used for.",
  "not_a_marker_of": "THE LEUKAEMIA-INITIATING COMPARTMENT IN THIS DISEASE. In infant MLL-rearranged ALL both CD34-positive/CD38-positive/CD19-positive and CD34-negative/CD19-positive cells initiate leukaemia on xenotransplantation [E1]; and in MLL-rearranged AML, LSCs derived from HSC origin and from CMP origin both carry a mature CD34-negative-to-low CD38-positive phenotype despite contrasting chemoresistance [E_q66_cell1 counter-evidence]. Functional origin and CD34 status come apart.",
  "how_shown": "knockdown and retroviral overexpression in primary human CD34-positive progenitors with clonogenic assay and expression profiling (PMID 18192237); prospective disease modelling from purified human HSC and CMP with humanised mouse xenografts and retrospective LSC deconvolution from patient samples (PMID 33627486); flow cytometry in patient cohorts (this entry's E1, E4, E7)",
  "nodes": [
    "cell-type-hsc-bone-marrow",
    "cell-type-pro-b",
    "cell-type-cmp",
    "cell-type-mep"
  ]
}
cell state
{
  "maintains_state": "NO. CD34 does not maintain stemness and should not be written as if it did. It is a surface glycoprotein whose loss shifts lineage proportions without abolishing progenitor function [E_q66_cell1], and in this disease the leukaemia-initiating state exists on both sides of the CD34 gate [E1].",
  "regulates_state": "Lineage-output bias in normal human haematopoiesis: CD34 promotes erythroid differentiation at the expense of granulocyte and megakaryocyte lineages [E_q66_cell1]. This is the only causal state claim CD34 supports anywhere in this entry, and it points at the erythro-megakaryocytic axis rather than at self-renewal.",
  "marks_state": "An immunophenotypic state that is clinically informative and mechanistically hollow: CD34 positivity at or above 10 percent in 82.9 percent of paediatric B-ALL [E7] and bimodal expression in adult KMT2A-rearranged pro-B ALL [E4]. Marking is not maintaining, and the graph has repeatedly used the former as if it were the latter.",
  "how_shown": "knockdown and overexpression with clonogenic lineage readout (PMID 18192237); xenotransplantation of sorted CD34-positive and CD34-negative fractions (this entry's E1)",
  "nodes": [
    "cell-state-stem-like-self-renewal"
  ]
}
How this could bite us back
  • Every CD34-selection experiment in this graph starts from a mixture. In patients with MLL-AF4, MLL-AF9 or MLL-ENL, the CD34+CD38-CD19-CD33- fraction is enriched for NORMAL HSCs while leukaemia-initiating cells sit in CD34+CD38+CD19+ and CD34-CD19+ fractions. Any experiment that reads a CD34-positive fraction as 'the leukaemia' or a CD34-negative fraction as 'not the leukaemia' is mis-specified in this disease.
  • CD34-directed depletion would not be lineage-restricted. CD34 is on vascular endothelium as well as haematopoietic progenitors, and it is the antigen used for autologous graft selection - a CD34-directed agent removes the rescue.
  • A CD34-anchored MRD gate will drift under therapy. CD34 is a differentiation-state marker and is downregulated during maturation; the menin-inhibitor immunophenotype study found dynamic phenotype shifts in 52% of patients, so any gate built on a stem-like marker needs re-definition per timepoint rather than per protocol.
  • The CD34+CD19-CD22+ escape reservoir argues for CD19/CD22 dual targeting, and lineage switch argues that dual B-lineage targeting is the wrong axis in KMT2A-rearranged disease. Project EVOLVE collected switches after CD19, CD22, CD38 and CD7 targeting. A dual B-antigen strategy may simply shorten the path to a myeloid relapse.
  • CD34 is expressed on endothelial cells, mast cells and hair-follicle stem cells as well as on haematopoietic progenitors (PMID 21499536). Any CD34-directed cytotoxic - antibody, ADC or CAR - is therefore a multi-tissue agent, and the vascular endothelium is the compartment a systemic agent reaches first.
  • Targeting CD34 would spare exactly the wrong cells in this disease: in infant MLL-rearranged ALL both CD34-positive and CD34-NEGATIVE fractions initiate leukaemia (this entry's E1), and in MLL-rearranged AML the LSCs from both cells of origin are CD34-negative-to-low (PMID 33627486). A CD34-directed agent would deplete the marrow's normal progenitor pool and leave the initiating population.
  • CD34 is not inert, so CD34 selection is not a neutral purification: CD34 biases lineage output, promoting erythroid at the expense of granulocyte and megakaryocyte differentiation (PMID 18192237). Immunomagnetic CD34 selection for transplant therefore enriches a population defined by a functional molecule, and cross-linking it with a bead-bound antibody is a perturbation nobody controls for.
  • The compartment itself is under-evidenced in the curated record: UniProt P28906 gives 'Membrane, single-pass type I' with no evidence code. Everything this graph assumes about CD34's accessibility comes from the fact that antibodies bind intact cells, which is operationally sufficient for a drug and is not a cell-biological measurement of where the protein is.
Still open
  1. Is the CD34+CD19-CD22+ escape reservoir present in infant KMT2A-rearranged B-ALL specifically, and at what frequency? The founding study is B-ALL at large and does not resolve the genotype.
  2. Does the CD34 status of the leukaemia-initiating fraction differ by KMT2A partner in a way that predicts lineage-switch risk? Aoki 2015 found partner-dependent answers and did not follow the patients to relapse phenotype.
  3. Are the CD34+CD38-CD19-CD33- cells in a t(4;11) infant's marrow truly fusion-negative, or do they carry the fusion without the secondary lesions? Polyclonal IGH and multilineage output do not exclude a pre-leukaemic carrier state, and this graph's genesis beat cares about exactly that distinction.
  4. Is there any evidence that the CD34 locus is a direct KMT2A-fusion target? None was found here, and the contrast with CSPG4 and PROM1, which are, is informative if it survives a dedicated search.
  5. Does the lineage-biasing function of CD34 operate in a KMT2A-rearranged blast? If CD34 pushes progenitors toward erythroid output, then bimodal CD34 in KMT2A-rearranged pro-B ALL may be marking a functional split rather than noise, and that is testable by sorting the two modes and comparing their transcriptomes.
  6. Is the CD34 intracellular tail phosphorylated in leukaemic blasts, and by which kinase? The mapped in vivo sites are not consensus PKC sites and one may be an AKT2 site (PMID 21499536), which would put CD34 downstream of a pathway this disease already cares about - and the mapping was done on the MOUSE protein, so the human sites are unknown.
  7. Nobody has ever imaged CD34 protein in a KMT2A-rearranged cell. Is it uniformly distributed on the surface, or polarised, as it is on endothelium? Polarity would matter for antibody and CAR engagement.
Every source behind this entry (10)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In a xenotransplantation study of primary infant MLL-rearranged leukaemia, CD34+CD38+CD19+ and CD34-CD19+ cells both initiated leukaemia in MLL-AF4 patients; CD34-CD19+ cells were the leukaemia-initiating cells in MLL-AF9 patients; and in MLL-ENL patients either CD34+ or CD34- cells were leukaemia-initiating depending on the pattern of CD34 expression. In patients with these MLL translocations, CD34+CD38-CD19-CD33- cells were enriched for normal haematopoietic stem cells with in vivo long-term multilineage repopulation capacity, repopulating recipients with B cells showing broad polyclonal IGH rearrangement and with CD4 single-positive, CD8 single-positive and double-positive T cells, whereas leukaemia-initiating cells developed leukaemic cells with clonal IGH rearrangement.

    PMID 25538041DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **Small numbers of primary patients per fusion**, not stated in the abstract. Xenograft leukaemia-initiating capacity in an immunodeficient mouse is an operational definition, not identity with the cell that sustains disease in a child; a fraction that fails to engraft may be niche-dependent rather than non-initiating. The normal-HSC conclusion rests on polyclonal IGH and multilineage output, which is strong but does not exclude a pre-leukaemic clone carrying the fusion without the secondary events.

    Evidence that argues against this (1)
    • That CD34 fractionation gives a different answer for MLL-AF4, MLL-AF9 and MLL-ENL in the same study is itself a warning that the result may be reading partner-specific differentiation stage rather than a property of CD34. A CRISPR-engineered human model comparing MLL fusion variants reports that partner identity governs leukaemia plasticity, which would produce exactly this pattern. Radszuweit P et al. 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In adult B-precursor ALL, NG2 expression was significantly associated with a CD10-negative, CD34-negative, CD24-negative, CD65s-positive, CD15-positive, CD13-negative, CD33-negative phenotype, and NG2 showed sensitivity 0.89, specificity 0.89 and positive predictive value 0.93 for MLL rearrangement.

    PMID 12886247DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **This is adult ALL.** The CD34-negative association is cited here specifically because it contradicts the paediatric convention, and the contradiction may simply be an age effect rather than an error in either source. Association, not a per-patient rule.

    Evidence that argues against this (1)
    • This graph's own `concept-flow-cytometry-immunophenotype-kmt2a-r` states that KMT2A-rearranged infant ALL is 'usually CD34-positive', citing a paediatric review. Both can be true across different ages, and the disagreement is the reason this entry refuses to give a single CD34 status for the disease. Gorecki M et al. 2023
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Among 1,382 newly diagnosed B-ALL cases, KMT2A-rearranged B-ALLs showed significantly lower mean fluorescence intensity of CD34, CD22, TdT and CD123 and significantly higher MFI of CD45, CD38 and CD58 than KMT2A-germline cases, and these differences held within the same EGIL immunological subgroup.

    PMID 38468826DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. Single-centre, mixed adult and paediatric. **MFI is a relative, instrument- and reagent-dependent quantity**, not an antigen density, so 'lower CD34 MFI' does not translate into a positivity threshold. The comparison group is KMT2A-germline B-ALL of all subtypes, so part of the difference is the genotype mix of the comparator.

    Evidence that argues against this (1)
    • A large iBFM study found that CD371-positive BCP-ALL - a different high-plasticity subgroup - showed strong expression of CD34 and CD45, so high CD34 MFI is not a general property of immature or plastic B-ALL and low CD34 MFI is not a general property of immature disease either. Buldini B et al. 2024
  4. E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    A 2026 Blood report describes cuplike nuclei and bimodal CD34 expression in adult KMT2A-rearranged pro-B acute lymphoblastic leukaemia.

    PMID 42720167DOI
    What this evidence cannot show

    **Title only.** The PubMed record carries no abstract, which is expected for a Blood images-in-haematology item. Routes attempted on 2026-09-20: Europe PMC core returned no PMCID for this identifier and PubMed efetch carried no `<ArticleId IdType="pmc">`, so routes 2-4 of the five-route test could not be attempted; the DOI route was not attempted because ashpublications is not open access. **This entry therefore asserts only what the title asserts** and does not say how many cases, in what proportion, or what 'bimodal' means quantitatively. Confidence weak per `concept-known-but-unreadable-literature`.

    Evidence that argues against this (1)
    • An images item is a single striking case or small series selected for being photogenic; it cannot establish that bimodality is characteristic. It is carried here only because it is a second, independent signal that CD34 in this genotype is not a clean binary. Bisegna ML et al. 2024
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CD22 expression precedes CD19 expression during B-cell development, and CD34+CD19-CD22+ cells are found in diagnostic and relapsed bone marrow samples of approximately 70% of patients with B-ALL. Their frequency increases twofold in patients in complete remission after CD19 CAR T-cell therapy, and the median frequency before treatment was threefold higher in patients who relapsed after CD19-directed immunotherapy over a median follow-up of 24 months. FISH on flow-sorted populations and xenograft modelling showed these cells harbour the genetic abnormalities present at diagnosis and initiate leukaemogenesis in vivo.

    PMID 35421218DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **B-ALL generally, not KMT2A-rearranged specifically** - the abstract does not resolve the genotype composition, so the threefold pre-treatment difference cannot be attributed to this graph's index disease. Detecting a rare CD34+CD19-CD22+ population reliably is a demanding flow measurement and the comparison is between groups defined retrospectively by relapse.

    Evidence that argues against this (1)
    • In KMT2A-rearranged disease the dominant escape from CD19-directed therapy is lineage switch to AML, not retention of a B-progenitor phenotype: 53 of 75 collected lineage switches were B-ALL to AML and 64.3% of the B-ALL-derived cases carried KMT2A rearrangement. A CD34+CD19-CD22+ reservoir may therefore be the wrong escape model for this genotype even if it is right for B-ALL at large. Silbert SK et al. 2025
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    CD34 was named in the text of 132 entries of this graph with no entry of its own, the top of a queue of 163 HGNC symbols in that state, ahead of CD33 at 30, CD38 at 21, PROM1 at 21, CD4 at 20, CSPG4 at 20, CD22 at 17, KIT at 13 and CD7 at 8.

    What this evidence cannot show

    Ranking reported by the coordinator from a run on disk on 2026-09-20 while other agents were writing, so it is a snapshot. Independently corroborated here by `grep -rl CD34 data/graph/nodes`, which returned **134** files - the two numbers differ because the grep counts files rather than entries and includes the coordinator's own edge and log files. **The count is of entries mentioning the symbol, not of mentions, and frequency is not importance.**

    Evidence that argues against this (1)
    • The same instrument ranked REST and EFS highly in a previous cycle, both of which are ordinary English words rather than the genes HGNC names. A rank from it is a reason to look and never a reason to write, and CD34 is written here because of what Aoki 2015 shows about it, not because it topped a list. Oncofusion knowledge graph (internal) 2026
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In 105 children with newly diagnosed B-ALL, 87 (82.9%) had bone marrow CD34-positive cells at or above 10% and 18 (17.1%) below. The CD34-negative group contained a significantly higher proportion of early pre-B-ALL and of E2A-PBX1 and MLL-AF4 fusion genes, and had significantly higher rates of post-induction MRD and relapse than the CD34-positive group, while no significant differences were found in CD10, CD19, CD20, CD22, CD79a, CD13, CD33 or CD38 positivity between the groups.

    PMID 38766795DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **Single-centre retrospective series of 105 children over two years with a CD34-negative group of only 18**, and a journal of limited standing. The prognostic conclusion is **confounded by construction**: the CD34-negative group is enriched for MLL-AF4 and TCF3-PBX1, both independently high-risk, so 'CD34-negative predicts relapse' and 'CD34-negative marks the high-risk genotypes' are not separated. Cited here for the composition finding, not for the prognostic claim.

    Evidence that argues against this (1)
    • Because the CD34-negative arm is enriched for MLL-AF4, this series cannot support a CD34 prognostic effect independent of genotype, and no multivariable adjustment for fusion status is reported in the abstract. The entry treats the prognostic claim as unsupported. Zha W et al. 2024
  8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    CD34 is UniProt P28906 (CD34_HUMAN), Hematopoietic progenitor cell antigen CD34, 385 amino acids, a single reviewed human record with primary gene name CD34 and no gene synonyms.

    What this evidence cannot show

    Fetched 2026-09-20 with `gene_exact:CD34 AND organism_id:9606 AND reviewed:true`, requesting `gene_primary` and `length` per rules 62 and 65, primary name read back, exactly one row returned. **Unlike most markers in this beat the gene symbol and the CD number coincide**, which is why this one is safe; CSPG4/NG2, IL3RA/CD123, MME/CD10 and PROM1/CD133 are not.

    Evidence that argues against this (1)
    • CD34 has a truncated splice isoform lacking most of the cytoplasmic tail, so the 385-residue canonical length is not the only CD34 a cell can display. Nothing about the accession bears on what a given flow antibody detects. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
  9. E_q66_loc1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CD34 is a type I transmembrane glycoprotein with a heavily glycosylated extracellular domain and a short signalling-competent intracellular tail: two tyrosine phosphorylation sites are reported on human CD34, the intracellular domain of mouse CD34 (residues 309-382) is phosphorylated by protein kinase C in vitro, and the in vivo sites mapped by mass spectrometry are NOT consensus PKC sites, with one potentially an AKT2 site - so kinases other than PKC act on the tail.

    CD34 Antigen: Determination of Specific Sites of Phosphorylation In Vitro and In Vivo. · Deterding LJ; Williams JG; Humble MM; Petrovich RM; Wei SJ; Trempus CS; Gates MB; Zhu F; Smart RC; Tennant RW; Tomer KB · Int J Mass Spectrom 2011
    PMID 21499536DOI
    What this evidence cannot show

    Mouse CD34 intracellular domain plus full-length mouse CD34 expressed in HEK293F cells - murine sequence in a human embryonic kidney host, not human haematopoietic cells and certainly not a KMT2A-rearranged blast. Residue numbering is for the mouse protein and must not be applied to human P28906 (385 aa) without re-mapping. The paper is about phosphorylation sites; the compartment statement is its framing sentence, not a measurement it made. Read: abstract only. Identifiers copied field by field from the efetch record fetched 2026-09-21, ArticleId anchored per brief rule 81.

    Evidence that argues against this (1)
    • UniProt annotates CD34's location as Membrane, single-pass type I membrane protein, with NO evidence code attached and no further compartment - no plasma-membrane-versus-endomembrane split, no apical or luminal statement, and no note. So the compartment that this graph targets with antibodies and CAR constructs rests, in the curated record, on an unattributed one-line annotation, and the experimental support has to come from the antibody literature rather than from cell biology. UniProt Knowledgebase, reviewed human record, fetched 2026-09-21 2026
  10. E_q66_cell1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CD34 is not a passive marker of the human haematopoietic progenitor - it BIASES lineage output: silencing CD34 in human HSCs enhances granulocyte and megakaryocyte differentiation and reduces erythroid maturation, with a matching gene-expression shift, and retroviral CD34 overexpression produces a marked increase in erythroid progenitors and a dramatic decrease in granulocyte progenitors by clonogenic assay - so CD34 promotes erythroid differentiation at the expense of the granulocyte and megakaryocyte lineages.

    Role of CD34 antigen in myeloid differentiation of human hematopoietic progenitor cells. · Salati S; Zini R; Bianchi E; Testa A; Mavilio F; Manfredini R; Ferrari S · Stem Cells 2008
    PMID 18192237DOI
    What this evidence cannot show

    Primary human CD34-positive progenitors with knockdown and retroviral overexpression, read out by clonogenic assay and expression profiling; a normal-haematopoiesis experiment with no leukaemia arm at all, and no KMT2A-rearranged cell. Overexpression from a retroviral vector is a supraphysiological perturbation. The effect is a shift in lineage proportions, not an identity switch in a single cell. Read: abstract only.

    Evidence that argues against this (1)
Connected to