IMPDH2 - the rate-limiting guanine-nucleotide enzyme whose inhibition lowers LEDGF and menin protein and dismantles the MLL-fusion complex itself
also called IMPDH2 · IMPD2 · IMP dehydrogenase 2 · inosine-5-prime-monophosphate dehydrogenase 2 · P12268
IMPDH2 is the bottleneck enzyme cells use to build one of the four letters of RNA and DNA, and leukaemia stem cells with a KMT2A fusion run it unusually hard. Blocking it does something unexpected: the cell makes less of the two scaffold proteins the fusion needs, so the cancer-driving machine falls apart and a menin inhibitor works better.
The same thing, for a scientist
IMPDH2 (UniProt P12268, 514 aa, primary gene name read back 2026-09-20) catalyses the rate-limiting oxidation of IMP to XMP in de novo guanine nucleotide biosynthesis. It is the best-evidenced single metabolic enzyme in KMT2A-rearranged leukaemia and the only one whose mechanism runs *back into* the fusion complex rather than merely alongside it. Leukaemia stem cells of MLL-rearranged AML show enhanced guanine nucleotide biosynthesis; targeting IMPDH2 reduces guanine nucleotides and rRNA transcription, which lowers the **protein** levels of LEDGF and menin, impairs formation and chromatin binding of the MLL-fusion complex, and reduces MLL target gene expression - and combining guanine-nucleotide or rRNA-transcription blockade with a menin inhibitor further suppresses MLLr AML [E1]. Independently, AMLs with MLL-fusions are susceptible to IMPDH inhibitors in vitro, and alternate-day dosing suppressed MLL-AF9-driven AML in vivo through overactivation of TLR-TRAF6-NF-kB signalling and VCAM1 upregulation [E2]. A third strand, not KMT2A-specific, shows guanine nucleotides sustaining ribosome biogenesis and suppressing the impaired-ribosome-biogenesis checkpoint, so IMPDH1/2 inhibition stabilises TP53 and represses MYC in AML LSCs regardless of TP53 mutational status [E3]. **The serious counter-result is that IMPDH-inhibitor sensitivity has been measured as independent of KMT2A status** in an AML panel, and absent in ALL lines [E4]. This entry therefore claims a strong, mechanistically explained *sensitisation* in KMT2A-rearranged myeloid disease, not a KMT2A-selective single-agent dependency.
This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.
I have never read one of these pictures before
A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.
Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.
Anything in yellow sticks is a bound small molecule, which is usually the drug.
Why this is the important entry in the metabolic layer
Most metabolic findings in this disease are parallel to the fusion: the blast has a metabolic state, and a drug attacks the state. IMPDH2 is different. The reported mechanism is guanine nucleotides -> rRNA transcription -> LEDGF and menin protein abundance -> MLL-fusion complex formation and chromatin binding -> MLL target genes 1. That places a metabolic enzyme upstream of the exact interface this project's central hypothesis is about, and it predicts synergy with menin inhibitors, which the same study reports 1.
It also inverts the usual framing of this beat's headline question. The metabolic phenotype here is not only a consequence of the fusion programme; the fusion complex is a consequence of metabolic supply. See concept-metabolic-dependency-consequence-or-independent-kmt2a.
IMPDH1 versus IMPDH2, and why the usual defence fails here
The two human IMPDH paralogues are a textbook identifier trap and they defeat the length check completely:
| accession | primary gene | synonyms | length |
|---|---|---|---|
| P12268 | IMPDH2 | IMPD2 | 514 |
| P20839 | IMPDH1 | IMPD1 | 514 |
Both are exactly 514 aa, and their recommended protein names differ by a single digit (Inosine-5-prime-monophosphate dehydrogenase 2 versus ...1). verify-accessions.py asserts accession plus length_aa and therefore cannot catch a swap between them - the same failure mode rule 68 documents for KRAS/NRAS/HRAS at 189 aa. Only gene_primary discriminates 5.
This matters because the papers differ in what they name. PMID 40102405 names IMPDH2 specifically as the target whose loss reduces LEDGF and menin 1. PMID 41191529 names IMPDH1/2 together 3. PMIDs 36453131, 41399259 and 41711315 name the enzyme family or the drug class (IMPDH, mycophenolic acid, ribavirin, mizoribine), which are pan-isoform, so those results are properly attributed to IMPDH activity, not to either gene 246. Mycophenolic acid inhibits both isoforms; attributing a mycophenolate result to IMPDH2 would be naming rather than measurement.
IMPDH1 has no KMT2A-specific evidence at all. Measured 2026-09-20 (Europe PMC REST; see ana-metabolic-enzyme-enumeration-kmt2a-bounded-absences for the instrument and its positive controls): IMPDH1 AND (MLL OR KMT2A) AND leukemia returns 36 records, of which 0 have IMPDH1 in title or abstract together with KMT2A/MLL in the title. The identical instrument returns 3 for IMPDH2, 2 for ACSL4 and 2 for MAT2A, so it is not a broken query. IMPDH1 is therefore recorded as a bounded absence here rather than given a thin entry of its own.
What the counter-evidence actually says, and why it is not fatal
Klootsema et al. set out to test whether IMPDH-inhibitor benefit was linked to KMT2A status and found it was not: sensitivity to IMPDH inhibitors was independent of KMT2A status; the synergy with cytarabine and fludarabine was SAMHD1-dependent rather than fusion-dependent; and the effect was absent in acute lymphoblastic leukaemia cell lines 4. Yague-Capilla et al. reach the same SAMHD1-centred conclusion with mycophenolic acid and ribavirin 6.
These do not contradict 1 or 2, which report a mechanistic consequence for the MLL-fusion complex and an in vivo effect in an MLL-AF9 model, not a comparative selectivity claim across genotypes. But they do bound the translational claim hard: an IMPDH inhibitor is not a KMT2A-selective agent. Its case in this disease is as a menin-inhibitor sensitiser and a nucleoside-analogue potentiator, and the lymphoid negative matters because the index disease of this graph is infant B-lineage.
Note also that 4 carries a title that names KMT2A-rearranged leukaemia while its finding is that KMT2A status does not matter. A title-only reading of this literature produces the opposite of the paper.
| uniprot | P12268 |
|---|---|
| length AA | 514 |
| alphafold ID | AF-P12268-F1 |
| ec number | 1.1.1.205 |
| pathway | de novo guanine nucleotide biosynthesis (rate-limiting IMP to XMP) |
| paralogue warning | IMPDH1 (P20839) is ALSO exactly 514 aa - length gives zero discrimination; only gene_primary separates them |
| resolution note | gene_exact:IMPDH2 AND organism_id:9606 AND reviewed:true returned 1 reviewed row on 2026-09-20; gene_primary read back as IMPDH2; length 514 aa asserted, with the explicit warning that the same assertion would pass for IMPDH1. |
| evidence lineage | myeloid (MLLr AML, MLL-AF9 mouse); ALL cell lines explicitly negative for the cytarabine-potentiation effect |
| direction in disease | pro-leukaemic supply function; inhibition is antileukaemic and sensitises to menin inhibition |
| subnuclear localization | damage-induced chromatin association, where IMPDH2 binds PARP1 and controls local NAD+ (PMID 39532854). NOT a body, not a speckle, not the nucleolus: no subnuclear structure has ever been assigned to IMPDH2, and the interesting spatial fact is the opposite of the usual one - a forced-nuclear IMPDH2 kills the cell, so the nuclear pool has to be kept small. |
| shown for fusion | NO. Measured 2026-09-21 with NCBI esearch: (IMPDH2[tiab] OR IMPDH[tiab] OR IMP dehydrogenase[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 1 record, title-checked off-topic (PMID 36843114, a KMT2A-rearranged AML phosphoproteomics study in which 'localization' is phosphosite localisation probability, not subcellular localisation). Operands: gene clause 1,477, gene AND fusion 6. 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). The gap is load-bearing here, because this entry's own E1 claims IMPDH2 inhibition dismantles the MLL-fusion complex and lowers LEDGF and menin protein - a claim about a nuclear assembly, resting on no measurement of where IMPDH2 is in such a cell. |
drugs pan isoform
[ "mycophenolic acid / mycophenolate mofetil", "mizoribine", "ribavirin" ]
localization
{
"subcellular": "CYTOSOL principally, with a real nuclear pool. UniProt P12268 annotates Cytoplasm, Cytoplasm cytosol and Nucleus, all at ECO:0000269. This is the only entry in this tranche whose default compartment is not the nucleus, and it changes how the protein should be read: IMPDH2 is a metabolic enzyme with a chromatin moonlighting job, not a chromatin protein with a metabolic sideline.",
"subnuclear": "CHROMATIN, and only on damage. IMPDH2 is enriched on chromatin in triple-negative breast cancer, that enrichment is DNA-damage dependent, and on chromatin IMPDH2 binds PARP1 and tunes its activity by controlling nuclear NAD+ availability. No nuclear body, speckle or nucleolar assignment exists. Forcing IMPDH2 to stay nuclear is lethal - nuclear NAD+ is depleted, PARP1 is cleaved and the cell dies - so the endogenous nuclear pool is small and transient by necessity.",
"how_shown": "subcellular and chromatin fractionation with proteomics, damage induction, PARP1 interaction and nuclear NAD+ measurement, plus a forced-nuclear IMPDH2 construct (PMID 39532854); UniProt P12268 SUBCELLULAR LOCATION comment.",
"systems": "triple-negative breast cancer cells; no haematopoietic and no KMT2A-rearranged system has ever been used to localise IMPDH2",
"node": "localization-chromatin-bound-versus-soluble-fraction-kmt2a"
}cell type
{
"expressed_in": "Not resolved to a haematopoietic cell type anywhere in this entry. IMPDH2 is broadly expressed; the leukaemia signal is about FLUX and demand, not about restricted expression, and the entry should not imply otherwise.",
"required_in": "LEUKAEMIA STEM CELLS of AML, by two independent routes. MLL-rearranged AML LSCs show enhanced guanine nucleotide biosynthesis and targeting IMPDH2 acts on them [E1]; separately, in human AML including venetoclax-resistant disease, IMPDH1/2 inhibition disrupts LSC stemness and suppresses reconstitution in xenotransplantation irrespective of venetoclax resistance or TP53 status [E_q66_cell1]. In vivo, alternate-day IMPDH inhibitor dosing suppresses development of MLL-fusion leukaemia [E2].",
"not_required_in": "Not selectively required in KMT2A-rearranged cells: in AML cell lines and primary AML samples, sensitivity to IMPDH inhibitors was INDEPENDENT of KMT2A status [E4]. The cell type that matters is the leukaemia stem cell, and the fusion is not what makes it sensitive.",
"how_shown": "metabolomics and flux measurement in sorted LSC, pan-isoform IMPDH inhibition, xenotransplantation reconstitution assays (PMID 41191529); in vivo alternate-day dosing in MLL-fusion leukaemia models (PMID 36453131)",
"nodes": [
"cell-type-hsc-bone-marrow",
"cell-type-gmp"
]
}cell state
{
"maintains_state": "LSC STEMNESS, measured as reconstitution capacity rather than as a marker phenotype: depleting guanine nucleotides disrupts LSC stemness and suppresses reconstitution of human AML in xenotransplantation [E_q66_cell1].",
"regulates_state": "THE GLYCOLYTIC ESCAPE FROM VENETOCLAX. Venetoclax inhibits OXPHOS; venetoclax-resistant LSC reprogramme to glycolysis; that shift is sustained by ribosome biogenesis running on de novo guanine nucleotides, and IMPDH1/2 inhibition blocks it, reactivating the impaired-ribosome-biogenesis checkpoint, stabilising TP53 and repressing MYC [E_q66_cell1]. This is a metabolic-state claim with a defined direction and a defined escape route, which is rare in this graph.",
"maintains_state_dna_damage": "Genome-maintenance competence, in a non-leukaemic system: IMPDH2 repression causes DNA damage accumulation and IMPDH2 tunes PARP1 through nuclear NAD+ [E_q66_loc1]. Recorded separately because it is a different cell type and a different readout from the LSC work, and pooling them would overstate both.",
"how_shown": "metabolic flux, checkpoint and TP53/MYC protein readouts in venetoclax-resistant human AML LSC with xenotransplantation (PMID 41191529); chromatin fractionation, damage assays and nuclear NAD+ measurement in TNBC (PMID 39532854)",
"nodes": [
"cell-state-stem-like-self-renewal",
"cell-state-hypoxic-glycolytic",
"cell-state-oxphos-dependence"
]
}- IMPDH inhibitors are approved immunosuppressants. Giving mycophenolate or mizoribine to an infant with leukaemia means deliberate immunosuppression in a patient already neutropenic from induction chemotherapy and at high risk of invasive fungal and viral infection; the in vivo result that spared immune function did so on an alternate-day schedule, which is a scheduling constraint, not a safety finding.
- Mycophenolate mofetil carries recognised reproductive toxicity and pregnancy-prevention requirements. That is not a barrier in infants but it is in adolescent and adult KMT2A-rearranged disease, where the same repurposing argument would be made.
- The mechanism runs through rRNA transcription and protein synthesis, the most broadly essential processes in the body. A therapeutic window that depends on leukaemia stem cells being more ribosome-hungry than gut crypt and marrow progenitors is a narrow window by construction.
- The counter-result is already published: IMPDH-inhibitor sensitivity was measured as independent of KMT2A status, so a trial enriched for KMT2A-rearranged patients on the strength of the mechanism could fail for reasons that were foreseeable in 2026.
- Guanine-nucleotide depletion raises TLR-NF-kB signalling in the reported mechanism. NF-kB activation is pro-leukaemic in other arms of this graph, so the therapeutic mechanism and a plausible resistance mechanism share a node.
- The paralogue defeats every attribution in this entry, and the length check gives no protection: IMPDH1 and IMPDH2 are both 514 aa, mycophenolic acid and ribavirin are pan-isoform, and the strongest leukaemia-stem-cell result deliberately writes IMPDH1/2 (PMID 41191529). Any claim that the target is IMPDH2 specifically is naming, not measurement - and IMPDH1 carries its own tissue liability, since IMPDH1 point mutants cause retinal disease (PMID 35013599).
- Forced nuclear localisation of IMPDH2 depletes nuclear NAD+, cleaves PARP1 and kills the cell (PMID 39532854). A tool or a drug that changes IMPDH2's compartment rather than its activity would therefore look like a potent cytotoxic with an unrelated mechanism, and the essential-function toxicity here is nuclear-NAD+ depletion in any dividing cell, not purine starvation.
- Selectivity: sensitivity to IMPDH inhibitors is independent of KMT2A status in AML lines and primary samples (this entry's E4). Repurposing mycophenolate here is therefore an AML-LSC strategy, not a KMT2A strategy, and a trial designed around fusion status would be designed around the wrong biomarker.
- Mycophenolate mofetil is an approved immunosuppressant given to transplant recipients and is used in GVHD prophylaxis. In an infant leukaemia population already lymphopenic from therapy, the on-target consequence of the repurposing is further immunosuppression, and no cited study measures infection risk as an endpoint.
- Does IMPDH2 blockade lower LEDGF and menin protein in infant KMT2A::AFF1 B-lineage cells, or is the LEDGF/menin effect myeloid-restricted like the cytarabine synergy?
- Is the reduction in LEDGF and menin protein specific, or a downstream consequence of globally reduced rRNA transcription and translation? A ribosome-agnostic way to deplete GTP would separate these.
- IMPDH1 versus IMPDH2: does the MLL-complex effect require IMPDH2 specifically, or would IMPDH1 loss do the same? No paper has tested the paralogues separately in this disease.
- If IMPDH-inhibitor sensitivity is genuinely independent of KMT2A status, what is the biomarker? SAMHD1 protein is the only one proposed, and it predicts nucleoside-analogue synergy rather than single-agent effect.
- Where is IMPDH2 in a KMT2A-rearranged blast, and does menin inhibition or IMPDH inhibition move it? This entry asserts that IMPDH inhibition dismantles the MLL-fusion complex and lowers LEDGF and menin protein - a nuclear-assembly claim with no accompanying measurement of IMPDH2's own compartment in that cell.
- Is the damage-induced chromatin pool of IMPDH2 present in leukaemia at all? If it is, the PARP1 link makes an IMPDH-inhibitor plus PARP-inhibitor combination mechanistically motivated rather than empirical, and this graph already holds a menin-PARP experiment to compare it against.
- Can any reagent separate IMPDH1 from IMPDH2 in a leukaemia cell? Until one is used, the entry's title names an isoform the experiments cannot resolve.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Leukaemia stem cells of MLL-rearranged AML exhibit enhanced guanine nucleotide biosynthesis; targeting IMPDH2 reduces guanine nucleotides and rRNA transcription, leading to reduced protein expression of LEDGF and menin, impaired formation and chromatin binding of the MLL-fusion complex, and reduced MLL target gene expression; inhibition of guanine nucleotide biosynthesis or of rRNA transcription further suppresses MLLr AML when combined with a menin inhibitor.
Guanine nucleotide biosynthesis blockade impairs MLL complex formation and sensitizes leukemias to menin inhibition. · Shi X, Li M, Liu Z, ..., Nakada D · Nat Commun 2025What this evidence cannot show
Abstract read this session; full text not retrieved, so the magnitude of the LEDGF and menin protein reduction, the antibodies used and the chromatin-binding assay were not checked. Myeloid (MLLr AML) system throughout; no B-lineage or infant KMT2A::AFF1 data are claimed by the abstract. The mechanism runs through rRNA transcription, which is a general growth axis, so a specific effect on the MLL complex has to be separated from a global reduction in protein synthesis - the abstract does not state which controls address this. Bounded absence: no independent replication of the LEDGF/menin protein result was found on 2026-09-20 in Europe PMC via IMPDH2 AND (MLL OR KMT2A) AND leukemia (32 records; 3 pass the title check and none is an independent replication).
Evidence that argues against this (1)
- A subsequent study designed specifically to test the KMT2A link found that sensitivity to IMPDH inhibitors was independent of KMT2A status in AML cells, that synergy with nucleoside analogues was SAMHD1-dependent rather than fusion-dependent, and that the effect was absent in acute lymphoblastic leukaemia cell lines. This does not refute the mechanism reported here but refutes any reading of it as a KMT2A-selective vulnerability. Klootsema Y et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Acute myeloid leukaemias with MLL-fusions are susceptible to IMPDH inhibitors in vitro; alternate-day administration of IMPDH inhibitors suppressed the development of MLL-AF9-driven AML in vivo without devastating immune function; mechanistically IMPDH inhibition induced overactivation of TLR-TRAF6-NF-kB signalling and upregulation of the adhesion molecule VCAM1, and combining IMPDH inhibitors with a TLR1/2 agonist effectively inhibited MLL-fusion AML development.
IMPDH inhibition activates TLR-VCAM1 pathway and suppresses the development of MLL-fusion leukemia. · Liu X, Sato N, Yabushita T, ..., Goyama S · EMBO Mol Med 2023What this evidence cannot show
Abstract read this session. First author is Liu X, not Kazama H - an earlier entry in this graph carried the wrong author string and it has been corrected; the string here was copied from the efetch record fetched 2026-09-20. The drugs used are pan-IMPDH (isoform-non-selective), so this result is evidence about IMPDH activity and cannot be assigned to IMPDH2 rather than IMPDH1. The in vivo claim rests on an alternate-day schedule chosen to spare immune function, which is a dosing-dependent result and not a general statement of tolerability. Mouse MLL-AF9; no infant B-lineage data. The proposed mechanism (TLR-NF-kB overactivation) is a gain-of-signalling explanation for a nucleotide-depleting drug and is indirect.
Evidence that argues against this (1)
- The accompanying commentary in the same issue frames the result explicitly as a repurposing proposal requiring clinical testing, noting that the antitumour effects of IMPDH inhibitors have not been proven in the clinical setting. Sezaki M et al. 2023
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In AML leukaemia stem cells, venetoclax-resistant LSCs undergo glycolytic reprogramming supported by enhanced ribosome biogenesis sustained by upregulated de novo guanine nucleotide biosynthesis; abundant guanine nucleotides suppress the impaired ribosome biogenesis checkpoint, destabilising TP53 and maintaining MYC expression. Inhibition of IMPDH1/2 depletes guanine nucleotides, activates the checkpoint, stabilises TP53, represses MYC, impairs the shift to glycolysis, disrupts LSC stemness and suppresses reconstitution in xenotransplantation, regardless of venetoclax resistance or TP53 mutational status.
Guanine nucleotides drive ribosome biogenesis and glycolytic reprogramming in acute myeloid leukemia stem cells. · Kawano G, Ikeda R, Ishihara D, ..., Kikushige Y · Blood 2026What this evidence cannot show
Abstract read this session. **This is an AML-wide LSC study, not a KMT2A study** - the abstract names no fusion and no KMT2A subgroup, so it is cited here for the guanine-nucleotide/ribosome/MYC axis and not as evidence in the index disease. It names IMPDH1/2 jointly and so does not discriminate the paralogues. The TP53-stabilisation arm is mechanistically attractive and correspondingly hard to rule out alternatives for from an abstract alone.
Evidence that argues against this (1)
- A contemporaneous study of IMPDH inhibition in leukaemia attributes the therapeutically relevant effect to depletion of the allosteric SAMHD1 activators GTP and dGTP and a consequent rise in active nucleoside-analogue triphosphates, not to ribosome biogenesis or MYC. Two different mechanisms are proposed for the same drug class in the same year and neither excludes the other. Yague-Capilla M et al. 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In AML cell lines and primary AML samples, sensitivity to IMPDH inhibitors was independent of KMT2A status; IMPDH inhibitors synergised with cytarabine and fludarabine in a SAMHD1-dependent manner in a subset of AML cells but not in acute lymphoblastic leukaemia cell lines; mechanistically IMPDH inhibition depleted the allosteric SAMHD1 activators GTP and dGTP, increasing active triphosphate metabolites in SAMHD1-proficient but not SAMHD1-deficient cells.
Targeting IMPDH to inhibit SAMHD1 in KMT2A-rearranged leukaemia. · Klootsema Y, Tsesmetzis N, Sharma S, ..., Herold N · Cell Cycle 2026What this evidence cannot show
Abstract read this session. Cell-line and primary-sample pharmacology; a negative selectivity result is only as strong as the panel's power, and the abstract does not give the number of KMT2A-rearranged versus wild-type lines tested, so this is an absence of a detected difference rather than a demonstrated equivalence. The ALL negative is in cell lines, not infant KMT2A::AFF1 primary material or PDX. The paper's title names KMT2A-rearranged leukaemia while its finding is that KMT2A status does not matter - read the abstract, not the title.
Evidence that argues against this (1)
- The earlier in vivo work reports suppression of MLL-AF9-driven AML development by IMPDH inhibitors in mice, an endpoint this cell-line study does not test; an absence of genotype-linked differential sensitivity in vitro does not exclude a genotype-linked effect on leukaemia initiation or stem-cell function in vivo. Liu X et al. 2023
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt reviewed human record P12268 (IMDH2_HUMAN) has primary gene name IMPDH2, synonym IMPD2 and length 514 aa; the paralogue P20839 (IMDH1_HUMAN) has primary gene name IMPDH1, synonym IMPD1, and is ALSO exactly 514 aa, so sequence length provides zero discrimination between the two.
UniProtKB entry P12268 (IMDH2_HUMAN), Inosine-5-prime-monophosphate dehydrogenase 2 · UniProt Consortium · UniProt 2026What this evidence cannot show
Resolved 2026-09-20 with gene_exact:IMPDH2 AND organism_id:9606 AND reviewed:true, fields including gene_primary and length; exactly one reviewed row returned and gene_primary read back as IMPDH2. The same query for IMPDH1 returned exactly one row, P20839, also 514 aa. A reviewed database record fixes the canonical sequence and symbol on the date read and says nothing about expression or function in KMT2A-rearranged blasts.
Evidence that argues against this (1)
- Asserting the accession and length is the standard defence in this graph, and it is explicitly insufficient here: because both paralogues are 514 aa, verify-accessions.py cannot detect an IMPDH1/IMPDH2 swap. The only protection is the primary gene name, exactly as rule 68 found for the 189-aa RAS family. Shi X et al. 2025
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The approved IMPDH inhibitors mycophenolic acid and ribavirin imbalance deoxyribonucleoside triphosphate pools and increase cytarabine efficacy in SAMHD1-proficient but not SAMHD1-deficient leukaemic cells.
IMPDH inhibition enhances cytarabine efficacy in SAMHD1-expressing leukaemia cells via guanine nucleotide depletion. · Yague-Capilla M, Dirks C, Eiden C, ..., Rudd SG · Mol Oncol 2026What this evidence cannot show
Abstract read this session. AML cell lines; the abstract makes no KMT2A claim at all, so this is cited for the drug mechanism and the SAMHD1 dependence, not for the index disease. Both drugs are pan-IMPDH and ribavirin has many documented activities beyond IMPDH, so a ribavirin result is weak evidence about IMPDH specifically.
Evidence that argues against this (1)
- The companion SAMHD1 study found the cytarabine-potentiating effect absent in acute lymphoblastic leukaemia cell lines, which bounds this mechanism away from the lymphoid index disease of this graph. Klootsema Y et al. 2026
- E_q66_loc1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
IMPDH2 has a chromatin pool and it is DNA-damage-dependent, not constitutive: in triple-negative breast cancer the de novo purine synthesis enzyme IMPDH2 is enriched ON CHROMATIN compared with other subtypes, IMPDH2 chromatin localisation is DNA-damage dependent, IMPDH2 repression leads to DNA damage accumulation, and on chromatin IMPDH2 interacts with and modulates PARP1 activity by controlling the nuclear availability of NAD+.
Nuclear IMPDH2 controls the DNA damage response by modulating PARP1 activity. · Espinar L; Garcia-Cao M; Schmidt A; Kourtis S; Gañez Zapater A; Aranda-Vallejo C; Ghose R; Garcia-Lopez L; Sheraj I; Pardo-Lorente N; Bantulà M; Pascual-Reguant L; Darai E; Guirola M; Montero J; Sdelci S · Nat Commun 2024What this evidence cannot show
Triple-negative breast cancer, not leukaemia and not a KMT2A-rearranged system - the claim is written with its system first for that reason (brief rule 73). The readout is chromatin enrichment by fractionation and proteomics, so it reports a fraction rather than a nuclear body. The most striking result in the paper is a PERTURBATION: when IMPDH2 is restricted to the nucleus it depletes nuclear NAD+, cleaves PARP1 and kills the cell, so a forced-nuclear IMPDH2 is not a model of the endogenous pool. 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)
- The compartment claim cannot be separated from the paralogue by any assay used in this literature. IMPDH1 and IMPDH2 are both 514 aa and are the two human isoforms of one enzyme; they assemble filaments with DIFFERENT architectures and different allosteric behaviour, and the leukaemia inhibitor work in this entry deliberately writes IMPDH1/2 together because mycophenolic acid and ribavirin do not distinguish them. So a 'nuclear IMPDH2' result obtained with an antibody or with a pan-isoform inhibitor is an isoform assignment by naming, not by measurement - the same shape as this graph's CSNK2A1 and RAS-family cautions. Burrell AL; Nie C; Said M; Simonet JC; Fernández-Justel D; Johnson MC; Quispe J; Buey RM; Peterson JR; Kollman JM 2022
- E_q66_cell1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In human AML leukaemia stem cells, guanine nucleotide supply IS the stemness state: venetoclax-resistant LSC undergo glycolytic reprogramming supported by enhanced ribosome biogenesis sustained by upregulated de novo guanine nucleotide biosynthesis, abundant guanine nucleotides suppress the impaired ribosome biogenesis checkpoint, destabilising TP53 and sustaining MYC - and IMPDH1/2 inhibition depletes guanine nucleotides, activates the checkpoint, stabilises TP53, represses MYC, blocks the glycolytic shift, disrupts LSC stemness and suppresses reconstitution of human AML in xenotransplantation, regardless of venetoclax resistance or TP53 mutational status.
Guanine nucleotides drive ribosome biogenesis and glycolytic reprogramming in acute myeloid leukemia stem cells. · Kawano G; Ikeda R; Ishihara D; Shima T; Sakoda T; Yamamoto S; Kochi Y; Semba Y; Ashitani S; Mori Y; Kato K; Maeda T; Miyamoto T; Soga T; Akashi K; Kikushige Y · Blood 2026What this evidence cannot show
Human AML, including primary cells and xenotransplantation, but AML and not infant B-lineage KMT2A::AFF1, and KMT2A status is not stated in the abstract - so this is a general AML-LSC result and the index disease is reached only by inference. The inhibition is pan-isoform (IMPDH1/2), so no IMPDH2-specific conclusion is licensed. The state readout is stemness by reconstitution capacity plus metabolic flux, which is stronger than a marker panel. Read: abstract only.
Evidence that argues against this (1)
- The KMT2A-selectivity of the axis is directly contradicted inside this entry: PMID 41399259 reports that in AML cell lines and primary AML samples, sensitivity to IMPDH inhibitors was INDEPENDENT of KMT2A status (this entry's E4). A cell-state mechanism that works in all AML LSC is a good drug rationale and a bad argument for a fusion-specific dependency, and both readings are currently in this entry. Cell Cycle 2026
- E_q66_depmapstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
IMPDH2 is ANTI-SELECTIVE in KMT2A-rearranged leukaemia lines, at the extreme end: on this project's genome-scale scan it ranks 17,563 of 17,728 - the 99.1st percentile - with a difference of +0.183, mean Chronos -0.684 in the 11 KMT2A-rearranged lines against -0.868 in 43 other leukaemia lines. The KMT2A-rearranged arm is LESS dependent on IMPDH2 than the comparator arm, not more.
Without a prior, which genes are most selectively required in KMT2A-rearranged leukemia lines? Genome-scale scan, 11 KMT2A-rearranged versus 43 other leukaemia lines, 17,728 genes · Oncofusion knowledge graph, analyses/depmap-kmt2a-selective-scan-full/results.json; DepMap Chronos via Open Targets 2026What this evidence cannot show
Eleven KMT2A-rearranged lines against 43 others is a small index arm, and the index arm is myeloid-enriched, so this number mixes fusion status with lineage (the lineage-split requirement in the standing brief). A difference of +0.183 on a pan-line mean of about -0.7 is modest in absolute terms; what makes it worth recording is its RANK, which is about as far from KMT2A-selective as the scan goes. This is a cell-line dependency measure and says nothing about leukaemia stem cells in vivo, where the supporting evidence in this entry sits.
Evidence that argues against this (1)
- The in vivo and LSC evidence in this entry is not refuted by this, because it is about a different quantity: E1, E2 and E_q66_cell1 are stem-cell, xenotransplantation and in vivo dosing results, and a Chronos screen in established lines cannot see a stemness or reconstitution phenotype. What the scan DOES refute is the fusion-specificity framing - and it agrees with this entry's own E4, which reports sensitivity to IMPDH inhibitors as independent of KMT2A status in AML lines and primary samples. Cell Cycle 2026
IMPDH2 blockade lowers LEDGF and menin protein and impairs MLL-fusion complex formation
guanine-nucleotide supply sets menin protein abundance via rRNA transcription
guanine-nucleotide supply sets LEDGF (PSIP1) protein abundance
guanine-nucleotide blockade sensitises MLLr AML to menin inhibition
MLLr AML leukaemia stem cells show enhanced guanine nucleotide biosynthesis
IMPDH1/2 inhibition represses MYC via the impaired-ribosome-biogenesis checkpoint in AML LSCs
IMPDH inhibition depletes the allosteric SAMHD1 activators GTP and dGTP
GTP/dGTP from the IMPDH pathway allosterically activate SAMHD1
approved pan-IMPDH inhibitors: mycophenolate, mizoribine, ribavirin
IMPDH2 was a positive control for the instrument; IMPDH1 is a bounded absence
IMPDH2's only spatial result is a chromatin-fraction enrichment that appears on DNA damage, in triple-negative breast cancer; forcing the enzyme to stay nuclear kills the cell, so the endogenous nuclear pool must be small and transient
venetoclax-resistant AML LSC reprogramme to glycolysis and the shift is sustained by de novo guanine nucleotides; IMPDH1/2 inhibition blocks the shift, reactivates the ribosome-biogenesis checkpoint, stabilises TP53 and represses MYC
the guanine-nucleotide axis is the documented ESCAPE from venetoclax-imposed OXPHOS inhibition, so it sits on the far side of this state rather than inside it
- regulates →The stem-like self-renewal state — a limited stem-cell module grafted onto a committed cell
IMPDH1/2 inhibition disrupts LSC stemness and suppresses reconstitution of human AML in xenotransplantation, regardless of venetoclax resistance or TP53 mutational status
guanine nucleotide depletion represses MYC through the impaired-ribosome-biogenesis checkpoint and TP53 stabilisation - a route to MYC that is mutation-independent and does not go through the menin-MLL axis