MAP3K14 (NIK) — the one node in this pathway where ACTIVATION, not inhibition, is the proposed therapy, and the one that also damages normal stem cells

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.

• NIK stabilisation impairs normal haematopoietic stem and progenitor self-renewal. The authors identify this as the central obstacle and propose a delivery solution they say does not exist.
• Constitutive non-canonical NF-kB activity is oncogenic in the B lineage - NIK is tumour-promoting in B-cell neoplasms and MAP3K14 variants activate NF-kB in mature B-cell malignancies. Proposing chronic NIK activation in an infant B-ALL would be proposing a known B-cell oncogenic signal as a therapy.
• NIK stabilisation represses canonical NF-kB, so it carries the infection risk of the canonical-blocking proposals in this folder as well as its own.
• The non-canonical arm builds secondary lymphoid organs and medullary thymic epithelium. Chronic activation in a developing child has developmental consequences no leukaemia experiment measures.
• TIFAB and HOXA9 each reverse the protective effect by downregulating RelB, so a resistance route to a NIK-based therapy already has two named genes and both are upregulated in this disease.

1. Does NIK stabilisation suppress or accelerate KMT2A-rearranged B-ALL? NIK is tumour-promoting in B-cell neoplasms and the only leukaemia data are myeloid, so the sign in the index disease is genuinely unknown rather than merely untested.
2. Is there any chemical matter that stabilises NIK? The proposal rests on an agent that does not exist; the cIAP-degrading SMAC mimetics stabilise NIK as a side effect and were not tested here.
3. Would NIK stabilisation combine with or cancel a menin inhibitor? Menin binds NFKB2, the non-canonical precursor, so a menin inhibitor may itself perturb the arm this proposal wants to activate.
4. Does verteporfin's effect survive a YAP/TEAD-independent control? Its RelB dependence is suggestive but its known pharmacology was not excluded.
5. Where does NIK-deltaT3 protein go once it is allowed to accumulate? The disease experiment in this entry is a stabilisation experiment, endogenous NIK is normally held near zero, and nobody has imaged the stabilised protein - so the compartment of the species that actually does the work in the published model is unknown.
6. UniProt's cytoplasmic call for NIK carries no evidence at all after three decades of NF-kB work. Is there a primary imaging paper for endogenous human NIK anywhere, and if not, is that neglect or is it that a protein held at near-zero abundance is effectively unimageable?
7. Should data/symbol-collisions.json carry a NIK row? Proposed to the coordinator with the measurement attached; the registry currently has none, and both candidate proteins share the same annotated compartment, which defeats content-based disambiguation.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Stabilisation of NIK using a non-degradable NIK-deltaT3 allele suppresses MLL-AF9-induced AML: it reduced colony formation and cell numbers in vitro, significantly delayed AML onset in vivo, delayed LSC repopulation, and inhibited LSC self-renewal on re-transplantation with no homing defect. Mechanistically NIK protein rose about two-fold with increased IKKa phosphorylation, RelB upregulation, p100-to-p52 processing, increased nuclear RelB and p52, reduced RelA expression and nuclear localisation, and upregulation of the canonical inhibitors Nfkbia and Tnfaip3. NIK also impairs the self-renewal of normal HSPCs, which the authors identify as the key obstacle, proposing LSC-restricted delivery via receptors such as TIM3 or CD93 as a currently unavailable solution. A Connectivity Map query on the NIK-stabilised LSC signature identified verteporfin, which dose-dependently suppressed leukaemic growth and colony formation, delayed AML in vivo, transiently upregulated non-canonical components, downregulated RelA protein, and was largely attenuated by RelB knockdown.

   Stabilization of NF-κB-Inducing Kinase Suppresses MLL-AF9-Induced Acute Myeloid Leukemia. · Xiu Y, Dong Q, Li Q, Li F, Borcherding N, Zhang W, Boyce B, Xue HH, Zhao C · Cell Reports 2018
   PMID 29320732DOI
   ▸What this evidence cannot show

   Genetic stabilisation of a mutant protein in mice, not a drug acting on endogenous NIK; there is no NIK stabiliser in existence. One fusion (MLL-AF9), one species, one laboratory. The human relevance claim rests on a supplementary figure. Verteporfin's selection is transcriptional-signature matching, and verteporfin has well-known YAP/TEAD and ROS activities that this design cannot exclude as the operative mechanism. Full text read this session via Europe PMC fullTextXML (PMC5810947), 44,387 characters.

   ▸Evidence that argues against this (1)

   • NIK is tumour-promoting in B-cell neoplasms, which the authors state as the contrast that made their result unexpected. Since the index disease of this graph is infant B-ALL, the direction that holds in myeloid AML cannot be assumed to hold in the disease this graph exists for. Li EW et al. 2026
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   Counted absence and enumeration. (KMT2A OR MLL OR "mixed-lineage leukemia") AND (MAP3K14 OR "NF-kappaB-inducing kinase"[tiab] OR "NF-kB-inducing kinase"[tiab]) returns 3 PubMed records: the MLL-AF9 NIK paper, a mantle cell lymphoma mutation meta-analysis, and a 2026 paper reporting that MAP3K14 synonymous variants drive altered splicing to activate NF-kB in mature B-cell malignancies. There is exactly one KMT2A-relevant NIK study in the literature.

   PubMed E-utilities esearch, db=pubmed, run 2026-09-20 · NCBI · PubMed 2026
   ▸What this evidence cannot show

   querytranslation was read. The short alias NIK was deliberately NOT searched as a bare token: it is a three-letter string that matches unrelated fields and would inflate the count, exactly the alias-precision failure the brief records for RD/NELFE. The 2026 Blood paper's abstract was fetched and read: two recurrent synonymous MAP3K14 variants introduce splice donor sites that remove the TRAF3-binding motif, and prime-edited cell lines show NF-kB activation with enhanced proliferation and survival across CLL, LPL/Waldenstroem, myeloma and mantle cell lymphoma. Those are MATURE B-cell malignancies, not precursor B-ALL, so the read-across to infant KMT2A-rearranged disease is an argument about the lineage's tolerance of non-canonical activation, not a result in it.

   ▸Evidence that argues against this (1)

   • A zero at the symbol level is not a zero at the pathway level, and on this beat that is demonstrable rather than hypothetical: PMID 29320732 studies the non-canonical arm directly in MLL-AF9 leukaemia and is unreachable from any of these subunit symbol queries, because it names the kinase and the pathway rather than the subunit. The count is honest about what was searched and should not be read as evidence that the biology is unstudied - only that it is not indexed under this symbol. Xiu Y et al. 2018
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   UniProt reviewed human record Q99558 (M3K14_HUMAN) has primary gene name MAP3K14 and length 947 aa; gene_exact:MAP3K14 AND organism_id:9606 AND reviewed:true returned exactly one row and gene_primary read back as MAP3K14.

   UniProtKB entry Q99558 (M3K14_HUMAN), Mitogen-activated protein kinase kinase kinase 14 · UniProt Consortium · UniProt 2026
   ▸What this evidence cannot show

   Resolved 2026-09-20 with gene_primary requested and all rows read. MAP3K7 (TAK1, O43318, 606 aa) was resolved the same way and returns zero relevant records in this disease; it is recorded as a bounded absence rather than given an entry.

   ▸Evidence that argues against this (1)

   • A reviewed database record fixes the canonical sequence and primary symbol on the date it was read, and nothing more. It does not establish that the protein is expressed in KMT2A-rearranged blasts, which isoform is present, or that the subunit acts in this disease. Accession and length are asserted so that a swap fails verify-accessions.py - a defence this graph needs because IKK-family and NF-kB subunit names collide heavily in the older literature - not because the record carries biological weight. Xiu Y et al. 2018
4. E_q89_map3k14_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   UniProt Q99558 (M3K14_HUMAN, primary gene MAP3K14, synonym NIK, 947 aa, read back 2026-09-21) carries exactly ONE subcellular location value, 'Cytoplasm', WITH AN EMPTY EVIDENCE LIST - no experimental, by-similarity or rule-based attribution of any kind - and no location note. Its TISSUE SPECIFICITY comment reads: 'Weakly expressed in testis, small intestine, spleen, thymus, peripheral blood leukocytes, prostate, ovary and colon.'

   UniProtKB Q99558 (M3K14_HUMAN) SUBCELLULAR LOCATION and TISSUE SPECIFICITY comments, fetched 2026-09-21 · UniProt Consortium · UniProtKB 2026
   ▸What this evidence cannot show

   Established as a statement about the record, and the record is close to empty on this dimension. An unevidenced location comment tells you what a curator believes, not what was measured, and this beat found no primary imaging paper for endogenous human NIK in eleven records of a live 146-record literature. The tissue comment is likewise uncoded, bulk-tissue and undated.

   ▸Evidence that argues against this (1)

   • The counter-case for the cytoplasmic call is strong on mechanism even though it is absent from the record, and the entry should not treat 'unsourced' as 'doubtful'. NIK's regulation is degradative and its destruction machinery - TRAF3, TRAF2, cIAP1/2 - is cytosolic; its substrate IKKa is cytosolic; and the whole non-canonical pathway is organised around keeping NIK out of existence until a receptor signal arrives. A protein consumed by a cytosolic complex is cytosolic. What the missing evidence costs is not confidence in the resting compartment but the ability to detect anything else: there is no baseline against which a NUCLEAR pool could be recognised if one appeared under stabilisation, and the disease-relevant experiment in this entry is precisely a stabilisation experiment. Nobody has asked where NIK-deltaT3 protein goes once it is allowed to accumulate, and that is a different question from where NIK sits when it barely exists. UniProt Consortium 2026
5. E_q89_map3k14_loc2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   'NIK' NAMES TWO DIFFERENT HUMAN KINASES AND BOTH ARE ANNOTATED CYTOPLASMIC. UniProt O95819 (M4K4_HUMAN, primary gene MAP4K4, 1,239 aa, read back 2026-09-21) lists gene synonyms HGK, KIAA0687 and NIK - for NCK-interacting kinase - and carries the subcellular location 'Cytoplasm' on ECO:0000269 PubMed:14966141. MAP3K14 (Q99558, 947 aa) carries the synonym NIK for NF-kappa-B-inducing kinase and the subcellular location 'Cytoplasm'. On 2026-09-21 the query NIK[tiab] AND ("MAP4K4"[tiab] OR "Nck-interacting kinase"[tiab]) returns 25 records. data/symbol-collisions.json was grepped before searching (93 rows at that moment) and re-grepped at hand-back after the coordinator grew it to 97 rows the same day; at both readings it contains no entry for NIK, MAP3K14 or MAP4K4.

   UniProtKB O95819 (M4K4_HUMAN) gene synonyms, SUBCELLULAR LOCATION and length, fetched 2026-09-21 · UniProt Consortium · UniProtKB 2026
   ▸What this evidence cannot show

   Verified live against both UniProt records this session. The 25-record co-occurrence query measures how often the two names appear together, which is a lower bound on contamination - papers using 'NIK' for MAP4K4 without ever writing MAP4K4 are not counted, and those are the dangerous ones. The registry absence was checked twice by grepping the file this session, at 93 rows and again at 97 rows after the coordinator committed four new rows mid-beat, and is stated as of 2026-09-21.

   ▸Evidence that argues against this (1)

   • One might argue the collision is harmless because context disambiguates - an NF-kB paper means MAP3K14 and a cytoskeletal or metabolic paper means MAP4K4 - and for most questions that is true. It is specifically NOT true for the question this beat was sent to answer. Context disambiguation works when the two proteins have different properties; here the property being looked up is IDENTICAL (both cytoplasmic), so a localisation search on 'NIK' returns confirmatory-looking records for either protein and the error is undetectable from the answer. The diagnostic run makes the size visible: the same localisation clause returns 11 records against MAP3K14 and 135 against NIK, a twelvefold inflation, and a beat that took the larger number as this gene's literature would have been reading mostly about a different kinase. That is the failure mode rule 65 describes for HDAC7/HDAC9, arriving through a shared alias rather than a shared protein name. UniProt Consortium 2026

• regulates →CHUK (IKK-alpha) — the only kinase that sits in BOTH NF-kB arms, which makes it the one subunit whose inhibition has an ambiguous sign in this disease

  NIK phosphorylates IKKa to drive p100 processing

• regulates →NFKB2 (p100/p52) — menin binds the non-canonical subunit too, and activating that arm SUPPRESSES MLL-AF9 leukaemia

  drives p100-to-p52 processing via IKKa

• regulates →RELB — the NF-kB subunit that SUPPRESSES MLL-AF9 leukaemia, with a direct ChIP target list and a genetic rescue, which makes it the best-evidenced anti-leukaemic node in this beat

  NIK stabilisation upregulates RelB and its nuclear localisation

• regulates →NFKBIA (IkB-alpha) — the switch itself. This graph has entries for everything the switch controls and none for the switch

  NIK stabilisation upregulates Nfkbia and Tnfaip3, lowering nuclear RelA

• inhibits →KMT2A-rearranged acute myeloid leukaemia in children and infants

  NIK stabilisation delays MLL-AF9 AML and reduces LSC self-renewal

• ← related toBlocking IKK/NF-kB in an infant is removing the switch every immune cell runs on — and the class has already failed clinically on exactly that

  the mirror risk: NIK stabilisation impairs normal HSPC self-renewal and mimics a B-cell oncogenic lesion

• related to →A partner's 'normal' compartment is usually read off a different tissue, and for at least two KMT2A partners that reading is wrong in blood cells

  NIK/MAP3K14's UniProt 'Cytoplasm' has an EMPTY evidence list - no experimental, by-similarity or rule support at all; and 'NIK' is shared with MAP4K4, which is also annotated cytoplasmic, so a NIK localisation search returns a right-looking wrong answer (Q89 2026-09-21)

• related to →The stem-like self-renewal state — a limited stem-cell module grafted onto a committed cell

  Stabilising NIK (NIK-deltaT3) reduced LSC self-renewal on re-transplantation in murine MLL-AF9 AML with no homing defect, and harms normal HSCs - the state is degradative stability, not differentiation (Q89 2026-09-21)

• ← related toJAK1 - recruited and stabilised by AF10 fusions, and the kinase whose downstream STAT1 the class I effect of menin inhibition explicitly does not need

  SECOND signalling kinase in this beat whose UniProt compartment has an EMPTY evidence list. MAP3K14's single 'Cytoplasm' value is unsourced (Q89); JAK1's single 'Endomembrane system' value is unsourced too, and worse, because JAK1's classical NLS and importin-alpha recognition were published in 2017 and the record carries no nucleus at all (Q93 2026-09-21)

• ← related toCHUK (IKK-alpha) — the only kinase that sits in BOTH NF-kB arms, which makes it the one subunit whose inhibition has an ambiguous sign in this disease

  The two-arm membership has a compartment consequence the graph should not blur: NIK's own cytoplasmic call is UNSOURCED and the nuclear step of the noncanonical arm belongs to RelB:p52, not to NIK (Q89) - whereas IKK-alpha genuinely shuttles, with a coded citation that is about its nuclear function (PMID 12789342) (Q93 2026-09-21)