Oncofusion

start / graph / gene-mthfd1

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q109-serine-one-carbon

MTHFD1 (P11586) - the inverse case: the deepest leukaemia dependency in the serine/one-carbon panel, and the drug does not touch it

also called C1-tetrahydrofolate synthase, cytoplasmic · C1TC_HUMAN · MTHFC · MTHFD

MTHFD1 is the big three-in-one folate enzyme in the cytoplasm. Of every gene in this beat, it is the one leukaemia cell lines actually need - and it is also the one a menin inhibitor leaves completely alone. That pairing is the clearest warning in the beat that a large drug-induced drop is not the same thing as a weak spot.

The same thing, for a scientist

MTHFD1 (P11586, 935 aa, exactly one reviewed human record, synonyms MTHFC and MTHFD) is the trifunctional cytoplasmic C1-tetrahydrofolate synthase, interconverting 5,10-methylene-THF, 5,10-methenyl-THF and 10-formyl-THF - the derivatives differentially required for purine and amino-acid biosynthesis [E3]. **It does not move under either drug:** RS4;11 day 7 VTP-50469 **+0.09**, EPZ5676 **+0.36**, from a detected baseline of 6.71 log2 CPM; day 2 -0.17; MOLM13 +0.06 from 7.05; and on the array the detected probe 202309_at is -0.07 in RS4;11, +0.13 in SEM and +0.07 in REH from baselines of ~11.05 log2 RMA [E1]. **And it is the deepest leukaemia dependency in the entire panel:** DepMap mean **-0.3512, 28.8% of 1,258 screens below -0.5, leukaemia mean -0.744 over 57 screens** - deeper in leukaemia than MEN1 (-0.5459) and DOT1L (-0.3931) in the same fetch [E2]. **So within one pre-registered panel, the genes that fall hardest (PHGDH 1.1%, ALDH1L2 0.0%, CBS 0.4%, CTH 0.0%) are the ones nothing needs, and the gene nothing moves is the one leukaemia lines need most.** Germline MTHFD1 loss causes folate-sensitive **neural tube defects** and a **combined immunodeficiency with megaloblastic anaemia** [E3].

Go deeper
The inverse relationship, stated as the panel's main safety finding

Within a single pre-registered panel, measured in one run 12:

geneRS4;11 d7 VTP% of 1,258 DepMap lines below -0.5leukaemia mean
CBS-4.070.4-0.0786
ALDH1L2-3.400.0-0.0207
PHGDH-3.241.1-0.0685
PSAT1-2.408.2-0.2669
CTH-1.570.0+0.1016
MTHFD1+0.0928.8-0.744
MAT2A (context)-0.4595.2-1.4653
DHFR (context)+0.2969.8-1.2688

The rank correlation between fold change and dependency in this panel is not merely weak, it is inverted. The three genes with the largest falls sit at 0.4%, 0.0% and 1.1%; CTH's mean gene effect is positive. The one-carbon sinks (MAT2A, DHFR, TYMS) are pan-essential and flat under the drug, and the one-carbon sources are drug-responsive and dispensable.

A coordinator reading this beat should take that table, not the fold changes, as its therapeutic content. No gene in this panel is nominated as a target on the basis of a menin-inhibitor fold change, and MTHFD1 is the reason the point is not merely rhetorical: the panel contains a real leukaemia dependency, and it is not one the drug produces.

A rule-91 worked example on the localisation record

P11586's only SUBCELLULAR LOCATION statement is Cytoplasm, with ECO:0000305 - curator inference, not experiment - citing PubMed 3053686 3. That paper is 'Primary structure of a human trifunctional enzyme. Isolation of a cDNA encoding methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydrolase...' (Hum DW, MacKenzie RE, J Biol Chem 1988) - a cDNA cloning and primary-structure paper.

So the citation supporting MTHFD1's compartment is a sequencing paper, and the code correctly says the compartment is an inference rather than a measurement. This is the cleanest instance in this beat of rule 91's point: the database row is a pointer, and here the pointer is honest about being an inference while still pointing at a paper that is not about localisation. PSPH's record has the same shape - ECO:0000305 citing a 1990 brain partial-purification paper.

The practical consequence for this beat: of eleven panel enzymes, three have no compartment at all (PHGDH, PSAT1, GPT2), two have an inferred one (MTHFD1, PSPH), one is uncoded (MTHFD2), and three are experimental (SHMT1, SHMT2, ALDH1L2, plus CBS). Any statement in this graph about a mitochondrial-versus-cytosolic split in the serine pathway rests on that mixture and should say so.

Facts and figures
gene symbolMTHFD1
uniprotP11586
length AA935
entry nameC1TC_HUMAN
protein nameC-1-tetrahydrofolate synthase, cytoplasmic
ensembl gene ID noteNOT ASSERTED. The live Ensembl resolution for MTHFD1 timed out during this beat's run and the script skipped the gene rather than guessing an id, so no ENSG is recorded here and MTHFD1 has no MV4;11 value. Recording the gap rather than supplying a plausible id is deliberate: this project has already had an Ensembl id resolve to the wrong gene (CTH to VSIG2).
accession resolutiongene_exact:MTHFD1 AND organism_id:9606 AND reviewed:true returns EXACTLY ONE reviewed row (rule 62 asserted): P11586, C1TC_HUMAN, gene_primary MTHFD1, synonyms MTHFC and MTHFD, 935 aa.
localizationCYTOPLASM, INFERRED NOT MEASURED. UniProt P11586's single SUBCELLULAR LOCATION statement is Cytoplasm with ECO:0000305 (curator inference) citing PubMed 3053686 - a cDNA primary-structure paper, not a localisation experiment. Recorded as an inference. No localisation experiment on MTHFD1 has been performed in any KMT2A-rearranged cell.
subcellular localizationCYTOSOLIC, by curator inference and by universal usage in the one-carbon literature, where MTHFD1 IS the cytosolic arm of the folate cycle opposite the mitochondrial MTHFD2/MTHFD2L. The annotation is ECO:0000305; the literature consensus is strong. Both are stated.
subnuclear localizationNONE, answered negative rather than left blank: no nuclear pool is annotated or reported for MTHFD1.
cell typeRS4;11 and SEM (KMT2A::AFF1 B-ALL), REH (KMT2A wild type) and MOLM13 (KMT2A::MLLT3 AML) measured here; NO MV4;11 value because the identifier resolution failed and was not guessed. Ubiquitously expressed per UniProt. The 57 DepMap leukaemia lines are the only functional evidence and they are not split by KMT2A status. NO primary infant sample and NO normal human pro-B cell.
cell stateMeasured in the drug-treated arresting state at day 7 of continuous menin-inhibitor exposure under uncontrolled medium folate supply - and it did not move in that state, which is the point. The state in which MTHFD1 loss has a human phenotype is DEVELOPMENTAL (neural tube closure) and HAEMATOPOIETIC/IMMUNE (megaloblastic anaemia with lymphopenia and combined immunodeficiency), the latter being directly relevant to a leukaemia patient's marrow and immune reconstitution.
epigenetic readoutNONE MEASURED at the MTHFD1 locus in any menin- or KMT2A-scoped study, and none expected to be informative: the gene does not move under either drug, so there is no transcriptional change for a chromatin measurement to explain. Recorded as a deliberate negative rather than as a gap. The wider epigenetic relevance of MTHFD1 is as a supplier of one-carbon units to the methionine cycle and therefore to every histone and DNA methylation reaction - and the pan-essential methionine-cycle enzyme MAT2A, which this graph already holds as gene-mat2a, sits immediately downstream and is also flat under the drug (-0.45).
shown for fusionNO - nobody has looked in any fusion. MTHFD1[tiab] AND the eight-operand fusion set returns exactly one record, an AML NADPH-metabolism paper read only by title. No KMT2A-fusion model, no menin-inhibitor experiment and no chromatin measurement exists for MTHFD1. Its transcript is flat in three lines measured here, which is a readout and not a fusion result.
mv411 independent labNOT MEASURED - the Ensembl symbol resolution timed out and the script skipped rather than guessing
absence classnever-measured
absence class noteMTHFD1 has never been measured in any KMT2A-fusion or menin-inhibitor experiment: the fusion-scoped literature is one unread AML paper, and this beat's own transcript null is the first such measurement. The DepMap leukaemia figure exists but has never been split by fusion status, which is a measured-not-analysed question sitting on top of a never-measured one.
the measurement rs411 d7 log2fc
{
  "vtp50469_330nM": 0.09,
  "epz5676_1uM": 0.36,
  "vtp_minus_epz": -0.26,
  "dmso_baseline_log2cpm": 6.71,
  "detected": true,
  "floor_collision": false,
  "d2_vtp": -0.171,
  "excluded_from_the_serine_arm_because": "|VTP| below the pre-declared 1.0 log2 threshold"
}
molm13 kmt2a mllt3
{
  "vtp": 0.062,
  "epz": -0.02,
  "dmso_baseline_log2cpm": 7.05,
  "detected": true
}
the measurement array gse63664 every probe
{
  "202309_at": {
    "rs411_dmso_log2_rma": 11.09,
    "rs411_mi22": -0.07,
    "sem_dmso": 11.1,
    "sem_mi22": 0.13,
    "reh_dmso": 11.04,
    "reh_mi22": 0.07
  },
  "239846_at": {
    "rs411_dmso_log2_rma": 5.92,
    "rs411_mi22": -0.2,
    "note": "BELOW the 6.0 log2 RMA floor in RS4;11; not interpreted"
  }
}
essentiality
{
  "depmap_release": null,
  "depmap_release_note": "Rule 83: no release identifier in the Open Targets payload. Fetched 2026-09-22.",
  "depmap_screens": 1258,
  "depmap_tissue_blocks_flattened": 29,
  "chronos_mean": -0.3512,
  "chronos_median": -0.2529,
  "pct_lines_below_minus_0_5": 28.8,
  "leukaemia_screens": 57,
  "leukaemia_mean_gene_effect": -0.744,
  "depmap_common_essential": "not asserted",
  "rank_in_this_panel": "DEEPEST leukaemia dependency of the eleven serine/one-carbon enzymes, and deeper in leukaemia than MEN1 (-0.5459) or DOT1L (-0.3931) in the same fetch - a lineage figure, NOT split by KMT2A status",
  "the_inverse_relationship": "MTHFD1 is the panel's most-required gene and its transcript is the flattest under both drugs; PHGDH, ALDH1L2, CBS and CTH fall hardest and sit at 1.1%, 0.0%, 0.4% and 0.0%."
}
How this could bite us back
  • **This is the entry that refuses the whole beat's temptation.** A large drug-induced fold change and a real dependency are anti-correlated across this panel, and MTHFD1 is the proof: nothing moves it and 28.8% of 1,258 lines need it, with a leukaemia mean of -0.744. Any target nomination arising from this beat must survive this table.
  • **Germline MTHFD1 loss causes a combined immunodeficiency with megaloblastic anaemia, lymphopenia and haemolytic uraemic syndrome, plus folate-sensitive neural tube defects.** Of every bite-back in this beat this is the one that lands on the organ system already limiting therapy: marrow and immune function in an infant receiving myelosuppressive chemotherapy.
  • **MTHFD1's dependency is a lineage figure, not a fusion figure.** The 57 leukaemia screens are not split by KMT2A status, so nothing here supports a fusion-selective claim, and a coordinator quoting -0.744 as a KMT2A dependency would be repeating exactly the error rule 84 and the lineage-split rule were written for.
  • **A flat transcript bounds transcription only.** Flux through a trifunctional enzyme can change with no expression change, which is the same argument this beat makes for ATF4 - so 'the drug does not touch MTHFD1' is a statement about mRNA and not about one-carbon supply.
Still open
  1. Is MTHFD1's leukaemia dependency KMT2A-selective? Its leukaemia mean of -0.744 over 57 screens has never been split by fusion status, and this project owns the scan that would do it (analyses/depmap-kmt2a-selective-scan-full). That is a compute-only question on data already held, and it would make MTHFD1 either the first real dependency this beat has found or a generic leukaemia requirement.
  2. Does one-carbon FLUX through MTHFD1 change under a menin inhibitor even though its transcript does not? The whole beat's ATF4 argument is that activity can collapse with an unmoved transcript, and the same logic applies here. A 13C-serine tracing experiment under VTP-50469 in RS4;11 would answer it for the whole pathway at once.
  3. Does the combined immunodeficiency and megaloblastic anaemia phenotype of MTHFD1 loss constrain any one-carbon intervention in this disease? A leukaemia patient's marrow and immune reconstitution are already the limiting toxicity of therapy, and this is the panel's only gene whose germline phenotype is haematopoietic and immune rather than neurodevelopmental.
  4. What does PMID 35385727 measure? It is the only fusion-scoped MTHFD1 record, it is about NADPH metabolism in AML, and it was left unread. Reading it is minutes and would either close this entry's bounded absence or confirm it.
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    On a pre-registered panel MTHFD1 does not move under either compound: RS4;11 day 7 VTP-50469 330 nM +0.09 and EPZ5676 1 uM +0.36 from a detected DMSO baseline of 6.71 log2 CPM, day 2 -0.17, MOLM13 day 7 +0.06 from a baseline of 7.05; and on the GPL570 array the detected probe 202309_at gives -0.07 in RS4;11 (baseline 11.09), +0.13 in SEM (11.10) and +0.07 in REH (11.04) while probe 239846_at is below the 6.0 log2 RMA floor in RS4;11 (5.92).

    What this evidence cannot show

    One line per platform, single doses, NO dose-response, n=2-3 per arm, day 7 of continuous exposure so arrest is a competing explanation. No series records the amino-acid composition of the medium, and this is an amino-acid-response panel. MI-2-2 in the array arm is the WEAK compound, so the REH wild-type contrast has little dynamic range. MTHFD1 has NO MV4;11 value in this run: the live Ensembl resolution for MTHFD1 timed out and the script SKIPPED the gene rather than guessing an id, so the independent-laboratory arm is missing for this gene alone. That is recorded as a gap and not filled.

    Evidence that argues against this (1)
    • A flat transcript is not evidence of a flat pathway. MTHFD1 is a trifunctional enzyme whose three activities are differentially required, and flux through it could change substantially with no transcript change at all - which is the same argument this beat makes about ATF4, where activity collapses with an unmoved transcript. The null here bounds transcription only. Ducker GS et al. 2017
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    MTHFD1 DepMap gene effect is mean -0.3512, median -0.2529, with 28.8% of 1,258 screens below -0.5 and a leukaemia mean of -0.744 over 57 screens - the deepest leukaemia dependency in this beat's pre-registered panel and deeper than MEN1 (-0.5459) and DOT1L (-0.3931) in the same fetch - with all 29 tissue blocks flattened and approvedSymbol read back as MTHFD1.

    What this evidence cannot show

    Rule 83: the payload carries NO release identifier, so depmap_release is recorded as null and absolute values drift quarterly; the cross-gene contrast is the durable part. Leukaemia screens are matched on the payload's own diseaseFromSource string against four substrings and are NOT split by KMT2A status. Rule 84: no median-split proportion is reported - the percentage is against an absolute -0.5 threshold. A CRISPR knockout in nutrient-replete medium understates a biosynthetic dependency. The leukaemia mean of -0.744 against a pan-lineage mean of -0.3512 is a LINEAGE enrichment and not a fusion enrichment: these 57 screens are not split by KMT2A status, so nothing here says MTHFD1 is more required in KMT2A-rearranged lines than in other leukaemias. Comparing it with MEN1 is legitimate only as stated - MEN1's shallower figure reflects a selective dependency and the comparison is about depth in leukaemia, not about importance.

    Evidence that argues against this (1)
    • A deep leukaemia dependency on a gene the drug does not move is an argument for MTHFD1 as an independent target and says nothing about menin biology - so this number, which is the most striking in the panel, is also the one least connected to this beat's subject. It could equally be cited to argue the whole serine beat is looking at the wrong genes. Pikman Y et al. 2016
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt reviewed human record P11586 (C1TC_HUMAN) has primary gene name MTHFD1 with synonyms MTHFC and MTHFD, length 935 aa, is a trifunctional enzyme interconverting 5,10-methylene-THF, 5,10-methenyl-THF and 10-formyl-THF (10-formyl-THF being required for purine biosynthesis), is ubiquitously expressed, carries a single SUBCELLULAR LOCATION statement of Cytoplasm with ECO:0000305 curator-inference support citing PubMed 3053686, and records three diseases: folate-sensitive neural tube defects, colorectal cancer, and combined immunodeficiency and megaloblastic anaemia with or without hyperhomocysteinaemia, whose manifestations include haemolytic uraemic syndrome, macrocytosis, epilepsy, hearing loss, retinopathy, mild intellectual disability and lymphopenia.

    What this evidence cannot show

    Resolved 2026-09-22 with gene_exact:MTHFD1 AND organism_id:9606 AND reviewed:true; exactly one reviewed row, gene_primary and length read back. Rule 91: the cited localisation paper PubMed 3053686 is a cDNA primary-structure paper, and the ECO:0000305 code correctly marks the compartment as an inference - the bibliographic record of that paper was verified but its abstract was not read. The colorectal-cancer disease association is a susceptibility association of a different evidential kind from the two Mendelian entries and should not be read as a monogenic phenotype.

    Evidence that argues against this (1)
    • The neural-tube-defect association is folate-SENSITIVE and derives largely from polymorphism epidemiology rather than from biallelic loss, and the only MTHFD1 leukaemia literature this beat found is of exactly that kind - a study of MTHFR and MTHFD1 polymorphisms and outcome in childhood ALL. Polymorphism associations are a weak basis for predicting the toxicity of pharmacological inhibition. Krajinovic M et al. 2004
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    PubMed 2026-09-22, one tag per term with translations read back: MTHFD1[tiab] returns 333 records; MTHFD1[tiab] AND the eight-operand KMT2A-fusion set returns 1, PMID 35385727 ('NADPH metabolism determines the leukemogenic capacity and drug resistance of AML cells', Cell Rep 2022), whose title was read and whose abstract was not; the alias MTHFC[tiab] returns 0 and MTHFD[tiab] returns 53 with 0 fusion-scoped; and MTHFD1[tiab] AND leukemia[tiab] returns 1, a study of MTHFR and MTHFD1 polymorphisms and outcome in childhood ALL.

    What this evidence cannot show

    Bounded by titles and abstracts only. The one fusion-scoped record was not read beyond its title, so whether it contains an MTHFD1 measurement in a KMT2A-rearranged line is unknown - that is a bounded rejection, not a bounded absence. The alias MTHFD returns 53 records which were NOT id-set-intersected with MTHFD1's 333, so only its fusion-scoped zero is used.

    Evidence that argues against this (1)
    • PMID 35385727 is an AML NADPH-metabolism paper and MTHFD1 is an NADPH-coupled folate enzyme, so this is precisely the record most likely to contain the measurement this entry says is missing - and it was left unread. Recorded as an unread on-topic paper rather than as an absence. Chen C et al. 2022
Connected to