Eleven immortalised lines passaged for decades: the competing explanation this beat could not close
also called cell line artefact · culture-acquired replication stress · DepMap line drift
Every number behind the Fanconi finding comes from cancer cells that have been growing in plastic dishes for twenty to forty years. Cells in long-term culture accumulate their own DNA-copying problems, so a repair dependency measured in them might be a property of the dish rather than of the leukaemia. Nobody has measured this in a child's actual leukaemia cells, and that is the gap this entry exists to name.
The same thing, for a scientist
The index arm is eleven Cellosaurus-annotated KMT2A-rearranged DepMap lines — MOLM-13, MOLM-14, MV4-11, NOMO-1, OCI-AML-2, SHI-1, THP-1, HB11;19, RS4;11, SEM, SEMK2 — every one an immortalised culture, several decades old, none infant-derived despite KMT2A::AFF1 being the defining infant lesion. Cancer cell lines undergo continuous genetic and transcriptional evolution in culture that measurably alters drug response, including between nominally identical stocks of the same line [E1]. Long-term culture selects for rapid proliferation, which is itself a source of replication stress, and for tolerance of the resulting genome instability — precisely the axis on which an FA dependency would be scored. **This objection is not answered by any analysis in this beat.** The line-label permutation controls which lines are the index arm but every line in both arms is an immortalised culture; the within-lineage contrast controls lineage but not culture history. The relevant evidence from primary material does not exist: a PubMed search for primary blasts in KMT2A-rearranged or MLL-rearranged leukaemia against DNA damage, γH2AX or 53BP1 returns **zero** records [E2]. What does exist is the capability — in vivo CRISPR screens in KMT2A-rearranged leukaemia and in leukaemic stem cells are published and routine [E3][E4] — so this is a gap in what has been done, not in what can be done.
Why this particular result is unusually exposed to it
Most DepMap dependencies would survive a culture objection because the biology being scored is not about proliferation. This one is not like that.
- The readout and the artefact share a mechanism. A CRISPR dropout screen measures proliferation over ~21 days. An FA-pathway dependency manifests as slowed proliferation under replication stress. A line that has been selected for decades to proliferate fast in culture has, by construction, been selected to tolerate or manage replication stress — and cells that manage it by upregulating fork protection would score as FA-dependent.
- None of the lines is infant-derived. KMT2A::AFF1 infant B-ALL is the disease this project exists for, and the four B-lineage lines (HB11;19, RS4;11, SEM, SEMK2) are childhood and adult ALL lines, not infant ones. The lineage half that carries the strongest FA signal is also the half furthest from the index disease in age of origin.
- Eleven lines, and several are close relatives. MOLM-13 and MOLM-14 derive from the same patient at different timepoints; SEM and SEMK2 are closely related. The effective independent n is smaller than eleven, which the line-label permutation does not correct.
- The comparator inherits the same problem. Forty-three other leukaemia lines, all immortalised. That helps — an artefact common to all cultured leukaemia lines cancels — but it only helps against a shared culture artefact, not against one that differs systematically between KMT2A-rearranged and other lines. KMT2A-rearranged lines are on average fast-growing and were often established from aggressive relapsed disease.
What would actually settle it, in rough order of cost
Cheapest, available today. Look for the FA genes in published in vivo CRISPR screens in KMT2A-rearranged leukaemia. At least two exist — combinatorial screens in KMT2A-rearranged leukaemia 3 and in vivo screening of leukaemic stem cells 4 — and if either carried FA genes in its library, the answer may already be sitting in a supplementary table. This beat did not do that search of supplementary data and it should be the next thing anyone does.
Moderate. Short-hairpin or CRISPR knockdown of FANCL or UBE2T in a KMT2A-rearranged patient-derived xenograft, read out as engraftment. PDX models for this disease and this technique are standard 3.
Definitive but expensive. Direct measurement of replication dynamics — DNA fibre analysis — in primary KMT2A-rearranged blasts against primary non-KMT2A blasts, which is the experiment in exp-replication-stress-readout-kmt2a-r.
Until one of those is done, the honest status of the FA finding is: a robust property of eleven cell lines, of unknown relevance to a patient. Every entry in this beat should be read with that sentence attached.
The one piece of human evidence, and which way it cuts
There is a single class of human observation bearing on this, and it is not encouraging for the dependency reading: KMT2A-rearranged leukaemia arises in people who have no functioning FA pathway at all. A child with Fanconi anaemia developed ALL, was treated with a topoisomerase-II inhibitor, and relapsed as AML carrying MLL-CBP t(11;16) 5; a second FA child evolved to MDS/AML with a t(9;11) and monoallelic KMT2A loss 6. Cryptic and recurrent lesions in FA-associated myelodysplasia and leukaemia have been catalogued systematically, including MLL partial tandem duplication 7.
That is a real human counterweight. It does not refute a quantitative fitness cost — DepMap measures exactly that, not viability — but it does exclude the strong reading, and it exists in patients rather than in plastic.
| kind | essential-function |
|---|---|
| severity | high — this is the one competing explanation this beat could not test |
| n infant derived lines | 0 |
| mitigation | search published in vivo and PDX CRISPR screens in KMT2A-rearranged leukaemia for FA genes before doing any new experiment; then knockdown of FANCL or UBE2T in a KMT2A-rearranged PDX read out on engraftment |
affects
[ "ana-fa-pathway-lineage-and-kmt2a-specificity", "ana-effect-size-vs-z-reranking" ]
index lines
[ "MOLM-13", "MOLM-14", "MV4-11", "NOMO-1", "OCI-AML-2", "SHI-1", "THP-1", "HB11;19", "RS4;11", "SEM", "SEMK2" ]
related line pairs
[ "MOLM-13 / MOLM-14 (same patient, different timepoints)", "SEM / SEMK2" ]
- If the dependency is culture-acquired, an FA-directed therapy would fail in patients while looking excellent in every preclinical cell-line experiment — the standard way this class of target wastes a decade.
- A replication-coupled dependency is weakest in quiescent cells, so even if it is real it may spare exactly the leukaemia-initiating population that causes relapse.
- Do FA genes appear in the libraries of the published in vivo CRISPR screens in KMT2A-rearranged leukaemia, and how did they score? This may be answerable from supplementary data alone.
- Is there any infant-derived KMT2A::AFF1 cell line in DepMap at all, and if not, what is the closest available model?
- Does the FA dependency weaken in quiescent leukaemia-initiating cells, as a replication-coupled mechanism would predict?
- How much of the eleven-line index arm's independence is lost to MOLM-13/MOLM-14 and SEM/SEMK2 relatedness?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Cancer cell lines undergo extensive genetic and transcriptional evolution during culture, to the extent that strains of the same line differ in drug response, so a dependency measured in long-passaged lines is a property of those cultures as much as of the tumour type.
Genetic and transcriptional evolution alters cancer cell line drug response. · Ben-David U, Siranosian B, et al. · Nature 2018What this evidence cannot show
A general result across cancer cell lines; it does not measure replication stress specifically, and it does not say that KMT2A-rearranged lines drift differently from other leukaemia lines.
Evidence that argues against this (1)
- Culture drift is shared by both arms of this comparison, so it cancels unless it differs systematically between KMT2A-rearranged and other leukaemia lines — which this paper does not establish and which remains an assumption in both directions. Oncofusion knowledge graph, analyses/fa-pathway-lineage-and-fusion-specificity 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No published study measures DNA damage burden in primary KMT2A-rearranged blasts: a PubMed title/abstract search for primary blasts in KMT2A-rearranged or MLL-rearranged leukaemia against DNA damage, γH2AX or 53BP1 returns zero records.
PubMed counted-absence search for DNA damage readouts in primary KMT2A-rearranged blasts, run 2026-09-20 · PubMed (NCBI) 2026What this evidence cannot show
A narrow conjunction of common words; γH2AX is often written in ways a tiab query will not match, and primary-sample DNA damage data frequently appear as one panel inside a paper about something else. This is an absence of a headline finding, not proof that the measurement has never been made.
Evidence that argues against this (1)
- Related searches do return substantive KMT2A-rearranged work on DNA repair and damage response in xenograft and primary-adjacent settings, so the field does study this disease with these tools; the specific comparison has simply not been published. Wright S et al. 2023
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Combinatorial CRISPR screening in KMT2A-rearranged leukaemia, including patient-derived material, is established and published — so testing the FA dependency outside immortalised lines is a matter of doing it, not of inventing a method.
Interrogating bromodomain inhibitor resistance in KMT2A-rearranged leukemia through combinatorial CRISPR screens. · Wright S, Hu J, et al. · Proc Natl Acad Sci U S A 2023What this evidence cannot show
This screen addresses bromodomain inhibitor resistance, not DNA repair, and whether FA genes were in its library is not stated in the abstract.
Evidence that argues against this (1)
- Availability of a method is not evidence about the result. It is equally possible that FA genes have been in in vivo libraries and simply did not score, which would be an unpublished negative and the most informative thing anyone could find. Vujovic A et al. 2023
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In vivo screening in leukaemic stem cells recovers dependencies that in vitro screens miss, establishing that the culture-versus-in-vivo distinction changes which genes score in this disease.
In Vivo Screening Unveils Pervasive RNA-Binding Protein Dependencies in Leukemic Stem Cells and Identifies ELAVL1 as a Therapeutic Target. · Vujovic A, de Rooij L, et al. · Blood Cancer Discov 2023What this evidence cannot show
RNA-binding proteins, not DNA repair, and leukaemic stem cells generally rather than KMT2A-rearranged disease specifically.
Evidence that argues against this (1)
- In vivo and in vitro differing does not say which direction an FA dependency would move. Quiescent leukaemic stem cells replicate less, so a replication-coupled dependency would be expected to weaken in vivo — which would make the DepMap result less, not more, relevant to relapse. Vujovic A et al. 2023
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A child with Fanconi anaemia developed early pre-B ALL and then, after topoisomerase-II-directed chemotherapy, a therapy-related AML carrying MLL-CBP t(11;16)(q23;p13) — a KMT2A-rearranged leukaemia in a host with no functioning FA pathway.
MLL-CBP fusion transcript in a therapy-related acute myeloid leukemia with the t(11;16)(q23;p13) which developed in an acute lymphoblastic leukemia patient with Fanconi anemia. · Sugita K, Taki T, et al. · Genes Chromosomes Cancer 2000What this evidence cannot show
One case; FA complementation group not stated in the abstract; therapy-related context with a topoisomerase-II inhibitor, which is the classic route to a KMT2A rearrangement and so a special case.
Evidence that argues against this (1)
- Occurrence is not fitness. A quantitative dependency of the size DepMap reports — a quarter to a half a Chronos unit — is entirely compatible with a clone forming and expanding in an FA-null host, just less efficiently. Quentin S et al. 2011
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A second Fanconi anaemia child evolved to MDS/AML with a der(9)t(9;11)(p24;q22) and monoallelic loss of ATM and KMT2A, associated with rapid leukaemic transformation and poor treatment response.
An uncommon t(9;11)(p24;q22) with monoallelic loss of ATM and KMT2A genes in a child with myelodysplastic syndrome/acute myeloid leukemia who evolved from Fanconi anemia. · Lovatel VL, de Souza DC, et al. · Mol Cytogenet 2018What this evidence cannot show
This is KMT2A deletion, not a KMT2A fusion, so it is a weaker analogy than E5 — it shows that the KMT2A locus is involved in FA-associated leukaemic evolution, not that a fusion oncoprotein operates without the FA pathway.
Evidence that argues against this (1)
- Monoallelic KMT2A loss and a KMT2A fusion are different lesions with different biology; citing this as evidence about fusion-driven leukaemia would be an overreach, and it is included here as context rather than as support. Sugita K et al. 2000
- E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Fanconi anaemia myelodysplasia and leukaemia carry a characteristic pattern of genomic abnormalities — 1q gain, 3q gain, -7/7q-, 11q- and cryptic RUNX1 lesions — with MLL partial tandem duplication among the rarer mutations found.
Myelodysplasia and leukemia of Fanconi anemia are associated with a specific pattern of genomic abnormalities that includes cryptic RUNX1/AML1 lesions. · Quentin S, Cuccuini W, et al. · Blood 2011What this evidence cannot show
57 FA patients; KMT2A rearrangement is not a feature of the recurrent pattern, and MLL-PTD is listed as rare.
Evidence that argues against this (1)
- Read the other way, this is mildly supportive of a dependency: if FA-null haematopoietic cells readily became KMT2A-rearranged leukaemias, that route should be common in this cohort, and it is not. Absence of KMT2A fusions from the FA leukaemia spectrum is weak evidence that the pathway helps. Sugita K et al. 2000
eleven immortalised lines, none infant-derived; the objection this beat could not close
eleven immortalised lines, none infant-derived; the objection this beat could not close
no analysis in this beat controls culture history; the line-label permutation redraws lines that are all immortalised cultures
tier 3 of the experiment, primary blasts, is the only arm that addresses this objection
SEM is microsatellite unstable, which the infant disease is not