The unnamed AML lines are NB4 and THP-1 - and NB4 is not KMT2A-rearranged at all
also called HOXA-AS2 cell lines · PMC7710717 provenance · lncRNA HOXA-AS2 AML system
An entry on the lncRNA HOXA-AS2 notes that the paper's AML cell lines are not named. The full text names two: NB4 and THP-1. Only THP-1 is KMT2A-rearranged; NB4 carries a completely different fusion, and it is the line the paper uses most.
The same thing, for a scientist
`noncoding-hoxa-as2-as3` carries the caveat *"The AML cell lines and patient numbers are not stated in the abstract"* for PMID 33268767 [E1]. The open-access full text names them by mention count: **NB4 (32), THP-1 (28)**, with single mentions of U937 and HL-60 [E2]. No GEO, SRA, BioSample or BioProject record is linked to the paper [E3]. The composition is the finding. **NB4 is a t(15;17) acute promyelocytic leukaemia line carrying PML::RARA and no KMT2A rearrangement**; THP-1 is t(9;11) KMT2A::MLLT3 monocytic AML. So the more-used line in a paper this graph cites in a KMT2A context is not a KMT2A leukaemia, and the HOXA-AS2 mechanism the entry describes - binding EZH2 and suppressing LATS2 - was established mostly in APL. HOXA-AS2 sits in the HOXA cluster, which is the programme KMT2A fusions sustain, so a reader naturally assumes a KMT2A context; the cell lines say otherwise for the majority of the work.
The two lines and what they are
| Line | Fusion | Disease | Mentions in full text |
|---|---|---|---|
| NB4 | PML::RARA, t(15;17) | acute promyelocytic leukaemia (M3) | 32 |
| THP-1 | KMT2A::MLLT3, t(9;11) | acute monocytic leukaemia (M5) | 28 |
| U937 | CALM-AF10 | histiocytic lymphoma / monocytic | 1 |
| HL-60 | none (MYC amplified) | AML M2 | 1 |
Only one of the four is KMT2A-rearranged, and the two main lines have different fusions, different lineages within myeloid, and very different HOXA-cluster biology: HOXA genes are characteristically high in KMT2A-rearranged and low in APL. A HOXA-AS2 function measured across both is a function measured across opposite HOXA states, which is either an unusually robust finding or a confounded one, and the deposition cannot tell them apart because there is no deposition 3.
The sibling caveat, and why it is a different class
The same entry's other flagged citation, PMID 30551418, carries "patient sample n and fusion genotypes are not stated". That one did not resolve: the paper has no PMCID in either PubMed or Europe PMC, no linked deposition, and a GEO keyword search on HOXA-AS2 returns 24 series of which every leukaemia-relevant one belongs to a different study - the rest are glioblastoma stem cells and microglia. It is recorded as a bounded failure in ana-unstated-systems-enumeration.
So one of the two HOXA-AS2 caveats resolves from the full text and the other resolves from nothing, which is a fair sample of this beat's overall yield.
| script | none - Europe PMC full text plus elink |
|---|---|
| status | run |
| result summary | Lines resolved to NB4 (PML::RARA APL, 32 mentions) and THP-1 (KMT2A::MLLT3, 28 mentions), with U937 and HL-60 once each. Only THP-1 is KMT2A-rearranged. No deposition exists for this paper, so the resolution rests entirely on full text. |
inputs
[ "PMID 33268767", "PMC7710717", "PMID 1995093", "PMID 6970727" ]
counts — the numbers behind the chart
{
"cell_lines_named": 4,
"kmt2a_rearranged_lines": 1,
"nb4_mentions": 32,
"thp1_mentions": 28,
"depositions": 0,
"fulltext_chars": 42462
}entries affected
[ "noncoding-hoxa-as2-as3" ]
import axes
[ "fusion: the dominant line carries PML::RARA, not a KMT2A fusion at all", "lineage: myeloid not B-lymphoid", "HOXA state: APL is HOXA-low, the opposite of KMT2A-rearranged disease" ]
- An lncRNA-directed strategy chosen from a mostly-APL dataset would be chosen in a HOXA-low context and applied to a HOXA-high disease, which is close to the worst-case mismatch for a HOXA-cluster noncoding element.
- HOXA-AS2 sits inside the HOXA cluster that KMT2A fusions occupy, so any oligonucleotide or CRISPR approach against it risks perturbing the coding HOXA genes the normal progenitor also needs.
- With no deposition, nothing in this citation can be re-analysed independently, so an entry built on it cannot be audited the way every other resolution in this beat can.
- Does HOXA-AS2 do anything in a KMT2A::AFF1 B-lineage cell, where the HOXA cluster is high and fusion-bound? The paper's KMT2A content is one myeloid line.
- Is a HOXA-AS2 function that holds in both HOXA-high (THP-1) and HOXA-low (NB4) contexts independent of HOXA output, and if so what is it actually doing?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The HOXA-AS2 paper reports that the long noncoding RNA functions as an oncogene by binding EZH2 and suppressing LATS2 in acute myeloid leukaemia, and its abstract does not name the AML cell lines used.
Long noncoding RNA HOXA-AS2 functions as an oncogene by binding to EZH2 and suppressing LATS2 in acute myeloid leukemia (AML). · Feng Y, Hu S, Li L, Peng X, Chen F · Cell Death Dis 2020What this evidence cannot show
Identifiers taken from the NCBI esummary record fetched on 2026-09-19. Full text read from Europe PMC PMC7710717 (open access) on the same date.
Evidence that argues against this (1)
- An lncRNA acting by binding EZH2 is a mechanism claim that has been made for a very large number of lncRNAs on the basis of RIP and knockdown alone, and the class has a poor replication record, so the mechanism deserves scepticism independent of which cell lines were used. Wang WT et al. 2019
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The paper's full text names NB4 32 times and THP-1 28 times, with single mentions of U937 and HL-60, identifying NB4 - a PML::RARA acute promyelocytic leukaemia line with no KMT2A rearrangement - as the most-used system.
Long noncoding RNA HOXA-AS2 functions as an oncogene by binding to EZH2 and suppressing LATS2 in acute myeloid leukemia (AML). · Feng Y, Hu S, Li L, Peng X, Chen F · Cell Death Dis 2020What this evidence cannot show
Counted by regular-expression tally over 42,462 characters of flattened Europe PMC full text on 2026-09-19. A mention count orders the lines by prominence and does not establish which figure used which, so 'most-used' is an inference from frequency. The paper deposited nothing, so no per-sample record exists to check the tally against - this is the one resolution in the beat with no repository corroboration at all.
Evidence that argues against this (1)
- A mention count can be inflated by a line appearing in every methods sentence of a small experiment and deflated by a line used once in a large one, so the ordering is suggestive rather than decisive and the underlying figures would settle it. Souren NY et al. 2022
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
NB4 is a maturation-inducible cell line carrying the t(15;17) marker, isolated from a human acute promyelocytic leukaemia, and THP-1 is a human acute monocytic leukaemia line; NCBI elink returns no GEO, SRA, BioSample or BioProject record for PMID 33268767.
NB4, a maturation inducible cell line with t(15;17) marker isolated from a human acute promyelocytic leukemia (M3). · Lanotte M, Martin-Thouvenin V, Najman S, Balerini P, Valensi F, Berger R · Blood 1991What this evidence cannot show
The NB4 founding report was fetched from PubMed on 2026-09-19 and establishes the line's origin and t(15;17) status; the THP-1 founding report (PMID 6970727) was fetched at the same time. The elink negative was run on 2026-09-19 and is informative for db=gds, which returned series for other papers in this beat, and weak for SRA, BioSample and BioProject, which returned zero for every paper tested including those with confirmed depositions.
Evidence that argues against this (1)
- THP-1's KMT2A::MLLT3 status makes it a legitimate KMT2A system, so the paper is not wholly off-target for this graph; the objection is to reading a two-line result as a KMT2A finding when the larger half is APL. Tsuchiya S et al. 1980
- related to →HOXA-AS2 and HOXA-AS3 — mid-cluster antisense transcripts with essentially no KMT2A-r evidence
the unnamed AML lines are NB4 and THP-1, and NB4 is not KMT2A-rearranged
- related to →KMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)
THP-1 is the only KMT2A-rearranged line in the paper
resolved source in the unstated-system worklist