SOCS2 (O14508) - a MENIN-INHIBITOR-SELECTIVE member of the signature, a cullin-5 E3 substrate receptor, and its likeliest upstream regulator is MEF2C rather than ATF4

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.

• SOCS2 is a brake on cytokine signalling and an E3 substrate receptor for EPOR and GHR. Lowering it should INCREASE erythropoietin and growth-hormone receptor signalling, which in a marrow under drug pressure is a growth input the leukaemia did not previously have - and no one has measured EPOR or GHR levels under a menin inhibitor.
• SOCS2 deficiency exhausts long-term HSC function on serial transplantation in mouse. A drug that lowers SOCS2 in the marrow is acting on a gene whose loss damages normal stem-cell durability, which in an infant with a lifetime of haematopoiesis ahead is the relevant toxicity axis.
• This is one of the few panel members where the drug's effect points in the therapeutically FAVOURABLE direction - lowering a poor-prognosis, stemness-promoting gene. That makes it the member most worth confirming functionally and the one most at risk of being over-read: the whole case rests on one cell line plus myeloid mouse models.

1. Is SOCS2's fall under a menin inhibitor MEF2C-dependent? A MEF2C degron or knockdown in RS4;11 that lowers SOCS2 without touching CHAC1 would move SOCS2 out of the ISR signature and into the menin-dependent target programme, where its behaviour already places it.
2. Why does SEM not express SOCS2 at all when RS4;11 expresses it at 11 log2 RMA? Both are t(4;11) KMT2A::AFF1 B-ALL lines in the same batch, and a 7 log2 difference between them is the largest inter-line discrepancy this beat found.
3. Does SOCS2 loss sensitise the leukaemia stem-cell compartment? SOCS2 promotes AML stem-cell quiescence, quiescence is what lets those cells survive chemotherapy, and a menin inhibitor lowers SOCS2 - but bulk cell lines contain no stem-cell compartment and the experiment requires a patient-derived xenograft.
4. Two of the four genes in a published 1,272-patient AML prognostic signature (SOCS2 and PHGDH) fall under a menin inhibitor. Is that mechanistic convergence or the generic behaviour of proliferation-correlated signature genes under any effective drug?

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

   SOCS2 falls 2.10, 1.65 and 1.58 log2 on three GPL570 probes under MI-2-2 in RS4;11 - the largest MI-2-2 effect in this beat's panel - with a clean REH control (-0.22, -0.34 at baselines of 7.79 and 9.77); it falls 1.06 log2 under VTP-50469 at day 7 while EPZ5676 gives +0.04, making it one of eight menin-inhibitor-selective genes in the 37-gene specificity table; it is NOT EXPRESSED in SEM (3.22 and 3.67 log2 RMA, at background); and in the same RS4;11 day-7 samples MEF2C falls 4.35 under VTP-50469 against 1.71 under EPZ5676 while CHAC1 falls 5.11 against 5.03.

   Q106 pre-registered ISR panel run over the cached GSE63664 (MI-2-2 / EPZ004777, SEM / RS4;11 / REH, GPL570 RMA) and GSE127506+GSE127508 (VTP-50469 330 nM / EPZ5676 1 uM, RS4;11 and MOLM13, d2 and d7, 3-prime-end RNA-seq) matrices. Script analyses/does-atf4s-own-transcript-move-under-a-menin-inhibitor/run.py, results.json, panel frozen in panel-preregistration.json BEFORE any expression value was read. SOCS2 probes 203372_s_at, 203373_at, 232539_at; RNA-seq row SOCS2; MEF2C, MEIS1, PBX3 and CHAC1 comparison rows. · NCBI GEO, re-analysed here 2026
   ▸What this evidence cannot show

   THE LOAD-BEARING CAVEAT: SEM does not express SOCS2 (both probes at background, 6-7 log2 below RS4;11), so SOCS2 is NOT a general property of the KMT2A::AFF1 fusion - it is a property of RS4;11, corroborated in MOLM13 (KMT2A::MLLT3, -1.38 at d2 and -1.76 at d7). The inter-line difference here is LARGER than the difference from the wild-type line. The third array probe's p-value is 0.12. RS4;11 array DMSO arm n=2. The menin-versus-DOT1L classification uses a 1 log2 threshold and SOCS2's difference is 1.10, so it clears it narrowly.

   ▸Evidence that argues against this (1)

   • A 1.10 log2 difference between two treated arms at a DMSO baseline of 5.61 log2 CPM is close to the threshold used to classify it, and reclassifying SOCS2 as shared rather than menin-selective would take only a small change in either arm. The menin-selectivity claim is the weakest of the eight so classified and should not be quoted without the effect sizes. NCBI GEO, re-analysed here 2026
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   High SOCS2 marks unfavourable acute myeloid and lymphoblastic leukaemias including those with MLL and BCR/ABL abnormalities and correlates with haematopoietic and leukaemic stemness signatures; SOCS2 and its correlated genes sit in regulatory networks fronted by IKZF1/Ikaros and MEF2C; lentiviral Mef2c transduction of murine haematopoietic precursors upregulates Socs2 while Ikzf1 transduction does not; and Socs2 deficiency unleashes HSC proliferation in vitro and exhausts long-term HSC function on serial bone-marrow transplantation.

   SOCS2 Controls Proliferation and Stemness of Hematopoietic Cells under Stress Conditions and Its Deregulation Marks Unfavorable Acute Leukemias. · Vitali C, Bassani C, Chiodoni C, Fellini E, et al. · Cancer Research 2015
   PMID 25858143DOI
   ▸What this evidence cannot show

   The MEF2C-to-SOCS2 causal test is MURINE, in haematopoietic precursors, by lentiviral OVEREXPRESSION - the right experiment in the wrong species and the wrong direction (gain, not loss). The human leukaemia association is expression-correlative across mixed AML and ALL cohorts, and 'MLL abnormalities' is not resolved into fusion partners, so KMT2A::AFF1 specifically is not addressed. The Ikzf1 negative control is what makes the MEF2C result usable rather than a bare correlation. Abstract read; figures not.

   ▸Evidence that argues against this (1)

   • If MEF2C drives SOCS2, then SOCS2's fall under a menin inhibitor is a second-order consequence of losing MEF2C and tells us nothing about SOCS2 itself - and MEF2C falls 4.35 log2 while SOCS2 falls only 1.06, which is a shallow transmission for a direct activator relationship. Either the relationship is indirect or other inputs dominate SOCS2 in this cell. NCBI GEO, re-analysed here 2026
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   SOCS2 is the top-scoring gene of a four-gene expression signature (SOCS2, IL2RA, NPDC1, PHGDH) that predicts overall survival as an independent prognostic parameter across AML cohorts totalling 1,272 patients and refines European LeukemiaNet risk classification, and in MLL-AF9 and Flt3-ITD/NPM1c mouse models SOCS2 promotes leukaemogenesis and increases the abundance, quiescence and activity of AML stem cells.

   SOCS2 is part of a highly prognostic 4-gene signature in AML and promotes disease aggressiveness. · Nguyen CH, Gluxam T, Schlerka A, Bauer K, et al. · Scientific Reports 2019
   PMID 31235852DOI
   ▸What this evidence cannot show

   AML ONLY - this is a myeloid signature and a myeloid mouse model set (MLL-AF9, Flt3-ITD/NPM1c), and the index disease here is B-lineage KMT2A::AFF1 (rule 73). The 1,272 figure is a total across several publicly available gene-expression cohorts, not a single prospective series, and signature discovery on public expression data with survival modelling carries a well-known overfitting exposure that the abstract does not quantify. Abstract read; figures not.

   ▸Evidence that argues against this (1)

   • Two of this signature's four genes - SOCS2 and PHGDH - fall under a menin inhibitor in this beat's data (-1.06 and -3.24). That could be a meaningful convergence or it could show that survival-associated gene signatures are enriched for proliferation-correlated transcripts that ANY effective cytostatic drug lowers, in which case the convergence carries no mechanistic information at all. Nothing here distinguishes them. NCBI GEO, re-analysed here 2026
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   UniProt O14508 is the single reviewed human SOCS2 record, 198 aa with an SH2 domain at 48-156 and a SOCS box at 151-197, synonyms CIS2, SSI2 and STATI2, thirteen PDB entries, annotated as the substrate-recognition component of a cullin-5 RING E3 ubiquitin ligase complex that mediates ubiquitination and proteasomal degradation of targets including EPOR and GHR; DepMap 26.06 mean gene effect +0.0016 across 1,258 screens, 0.1% of lines below -0.5, leukaemia mean -0.0131.

   UniProtKB - O14508, SOCS2_HUMAN, suppressor of cytokine signaling 2 · UniProt Knowledgebase 2026
   ▸What this evidence cannot show

   Resolved live 2026-09-22: exactly one reviewed record (rule 62 satisfied), re-fetched alone with name, length and primary symbol compared back: PASS. RULE 91 FLAG: the Cytoplasm location's ECO:0000269 points at PubMed 9727029, a yeast-two-hybrid paper on SOCS2 binding the IGF-I receptor - not a localisation study, so the compartment call is a by-product of an interaction paper. The thirteen PDB entries were NOT fetched and nothing is claimed from them. DepMap figures from the same session with the release read back (rule 83).

   ▸Evidence that argues against this (1)

   • A flat DepMap gene effect on a gene that two independent papers report as promoting leukaemogenesis and stem-cell quiescence is a direct contradiction, and the likeliest resolution is that neither property is measurable in a two-week pooled proliferation screen - quiescence in particular is invisible to an assay that selects for growth. But the contradiction should be stated rather than explained away: the functional genomics do not support the functional claims. DepMap / Open Targets Platform 2026
5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   PubMed returns 2 records for SOCS2 jointly with an eight-operand KMT2A-fusion list - PMID 31235852 and PMID 25858143, both title-read and both genuinely about SOCS2 in leukaemia including MLL-abnormal disease - and 0 for SOCS2 with a thirteen-name menin-inhibitor list with a non-empty querytranslation and 'No items found.'; the alias CIS2 reaches 680 records sharing only 1 with SOCS2's 849 (Jaccard 0.0007) and STATI2 reaches zero records in the whole index.

   NCBI PubMed esearch, run 2026-09-22, with querytranslation and warninglist read back from the response for every query quoted · NCBI PubMed 2026
   ▸What this evidence cannot show

   Both on-topic titles were fetched and read and both abstracts were obtained, so these two are evidence and not leads. Rule 3 satisfied on the menin-drug zero. THE CIS2 MEASUREMENT IS NOTABLE: 680 records sharing ONE id with SOCS2 means CIS2 is effectively a different entity in the literature (largely unrelated usages), so widening a SOCS2 query with CIS2 would import 679 off-topic records - the same shape as the CHOP trap. STATI2 returns zero records and is not a usable token. Rule 94 form; rule 95 observed; rule 80: titles and abstracts only.

   ▸Evidence that argues against this (1)

   • Two fusion-scoped records is a thin literature for a gene with 849 total records, and neither addresses KMT2A::AFF1 or B-lineage disease specifically - both are AML-centred. The SOCS2 story in this graph therefore rests on myeloid evidence imported into a B-lineage question, which is this project's single most frequently documented synthesis failure. NCBI PubMed 2026

• ← regulatesMEF2C — an output-program transcription factor, the second-strongest leukemia co-dependency, and a gene humans cannot spare one copy of

  lentiviral Mef2c transduction upregulates Socs2 in murine haematopoietic precursors while Ikzf1 does not (PMID 25858143); MEF2C falls 4.35 log2 under VTP-50469 and only 1.71 under EPZ5676, which explains SOCS2's menin-selectivity without the ISR

• related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)

  largest MI-2-2 effect in the panel (-2.10 on three agreeing probes) in RS4;11 with a clean REH control - but SEM does not express SOCS2 at all, so this is an RS4;11 property, not a general KMT2A::AFF1 one

• ← related toRevumenib (SNDX-5613, Revuforj) — first approved menin–KMT2A interface inhibitor

  class effect: SOCS2 is MENIN-INHIBITOR-SELECTIVE (VTP-50469 -1.06, EPZ5676 +0.04). Measured for VTP-50469 and MI-2-2, not for revumenib itself

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

  top gene of a four-gene AML prognostic signature across 1,272 patients, and promotes leukaemogenesis and AML stem-cell quiescence in MLL-AF9 and Flt3-ITD/NPM1c mouse models (PMID 31235852)

• related to →IKZF1 - the defining co-lesion of every other high-risk B-ALL, and this disease does not use it until after treatment

  SOCS2-correlated genes sit in networks fronted by IKZF1/Ikaros and MEF2C, but Ikzf1 transduction does NOT upregulate Socs2 - the negative control that makes the MEF2C result usable (PMID 25858143)

• related to →The differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia

  SOCS2 increases the abundance, quiescence and activity of AML stem cells; this beat measured it only in bulk drug-arrested cell lines, which contain no stem-cell compartment

• ← supportsATF4's own transcript does NOT move under either menin inhibitor - and the whole ISR arm is reproduced by a DOT1L inhibitor, so it is not a menin-interface readout

  SOCS2 is one of eight menin-selective genes (VTP -1.06, EPZ5676 +0.04) and is not expressed in SEM at all (3.22, 3.67 log2 RMA)