CXXC4 (IDAX) - TET2's CXXC domain, sitting on a separate gene after an evolutionary inversion, and annotated cytoplasmic only by similarity
also called CXXC4 · IDAX · Inhibition of the Dvl and axin complex protein · CXXC-type zinc finger protein 4
CXXC4 is a tiny protein that is essentially a spare address-reader: it is the CG-reading module that TET2 used to carry, split off onto its own gene early in evolution, and it still delivers TET2 to unmethylated CG islands. Whether it is even in the nucleus of a human cell has never actually been measured - the database says cytoplasm, by guesswork from other species.
The same thing, for a scientist
CXXC4 (Q9H2H0) is **198 amino acids**, of which the CXXC-type zinc finger at **132-173** is a fifth [E1]. It is the smallest member of the family by a factor of three and is close to being a bare reader domain with a linker. Ko et al. established what it is for: **IDAX/CXXC4 was originally encoded within an ancestral TET2 gene that underwent a chromosomal inversion separating the TET2 CXXC domain from the catalytic domain**, and the IDAX CXXC domain binds unmethylated CpG, localises to promoters and CpG islands, and interacts directly with the TET2 catalytic domain [E2]. It therefore supplies in trans the address module that `gene-tet2` lacks in cis. The functional consequence they report is counter-intuitive and worth flagging: IDAX expression causes **caspase activation and TET2 protein down-regulation**, dependent on DNA binding through the IDAX CXXC - so the reader recruits TET2 to DNA and then precedes its destruction [E2]. **Rule-91 point, and it is the strongest in this beat:** UniProt's only subcellular-location term for CXXC4 is **Cytoplasm, ECO:0000250 - by similarity** [E1]. There is no experimental localisation annotation for human CXXC4 at all, and the one annotation that exists places it outside the nucleus. A protein that may not be nuclear cannot be assumed to compete for a chromatin address. Its other documented role is as a negative regulator of Wnt signalling via DVL1 [E1], and in embryonic stem cells Idax and its paralogue Rinf/CXXC5 together facilitate expression of TET enzymes during differentiation [E3]. The only haematological data is correlative: CXXC4 mRNA is lower in 154 MDS patients than in 50 controls, higher in advanced WHO subtypes, and associated with overall survival on univariate but **not** multivariate analysis [E4]. In DepMap it is flat (difference -0.002, rank 8,853 of 17,728) [E5].
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.
A reader domain that became its own gene
The inversion story is the reason this entry exists. In an ancestral genome the CXXC domain and the catalytic domain were one TET2 gene; a chromosomal inversion split them, and the human genome now carries CXXC4 adjacent to TET2 on chromosome 4 2. Liu et al. note the parallel arrangement at TET3/Cxxc10-1 and group the CXXC domains of CXXC4, CXXC5, TET1 and TET3 into one homology subfamily 6.
Two consequences for this beat:
- TET2 is not a CXXC protein but it is not addressless either. The census excludes it correctly, and
gene-tet2says why; CXXC4 is where the excluded domain went. - A trans-encoded reader is dosage-separable from its enzyme in a way a cis-encoded one is not. CXXC4 expression can rise or fall independently of TET2, which is exactly what the MDS correlation measures 4 and is a class of regulation no cis-domain protein has.
The localisation problem, stated as a defect in the record and not in the protein
UniProt lists exactly one subcellular location for human CXXC4: Cytoplasm, with evidence code ECO:0000250 - inferred from a homologue, not measured 1. There is no nuclear annotation, coded or uncoded, and no isoform-scoped statement.
That is awkward in both directions:
- The functional literature has IDAX localising to promoters and CpG islands in genomic DNA 2, which is a nuclear statement, and has Idax regulating TET expression in differentiating ES cells 3. The annotation and the literature disagree.
- If a beat in this project needed a 'where is this protein' field, the honest answer is unknown for the human protein, and this entry records it that way rather than borrowing the functional literature's implication.
This is a third distinct shape of the rule-91 problem, alongside TET1 (experimental codes present but isoform-scoped) and TET3 (three locations, all by similarity): one location, by similarity, and it contradicts the functional literature. Recorded as a correction owed to curation rather than as a finding.
What can be said about blood
Very little, and this section exists so nobody overstates it later.
CXXC4[tiab]returns 51 records andIDAX[tiab]returns 16; the IDAX hits were title-checked and are a mix of CXXC4 papers, Mullerian-inhibiting-substance papers and one chemistry paper, so IDAX is usable but needs checking 7.- The only haematological dataset is 154 MDS marrows and 50 controls, by qRT-PCR: CXXC4 and TET2 mRNA both significantly lower in MDS; CXXC4 higher in RAEB-1/RAEB-2 than in early subtypes; CXXC4 associated with overall survival but not leukaemia-free survival, and not independent on multivariate analysis, where only age, haemoglobin and marrow blasts survived 4.
- Nothing in KMT2A-rearranged leukaemia. The string CXXC4 does not appear in the full text of the index infant whole-genome study 8.
So CXXC4 is written here as a family member with a clear molecular job and no established role in this disease. That is the honest state and the entry should not be read as more.
| uniprot | Q9H2H0 |
|---|---|
| gene symbol | CXXC4 |
| protein name | CXXC-type zinc finger protein 4 |
| length AA | 198 |
| alphafold ID | AF-Q9H2H0-F1 |
| accession resolution | UniProt REST gene_exact:CXXC4 AND organism_id:9606 AND reviewed:true, run 2026-09-21; one row; primary gene name read back as CXXC4 and length 198 from the same fetched record. |
| alias hazard | IDAX[tiab] returns 16 records including Mullerian-inhibiting-substance and computational-chemistry papers; usable but every title must be read. CXXC4[tiab] returns 51. |
| function | in-trans CXXC reader for TET2: binds unmethylated CpG, localises to CpG islands, binds the TET2 catalytic domain directly; separately a negative regulator of Wnt signalling via DVL1 |
| classification | the orphaned TET2 CXXC domain, now its own gene; the family's smallest member and the clearest case of a reader separable in dosage from its enzyme |
| relationship to tet2 | trans-encoded address module; the ancestral TET2 gene's CXXC after a chromosomal inversion |
| shown for fusion | NO - nobody has looked in any fusion |
| cell type evidence | mouse embryonic stem cells and embryoid bodies; human U937 monocytic line; 154 MDS marrows by qRT-PCR. No KMT2A-rearranged material of any kind. |
| subnuclear localization | Unknown for the human protein. UniProt's only location term is Cytoplasm, ECO:0000250 (by similarity), with no experimental annotation - and the functional literature places IDAX at promoters and CpG islands, contradicting it. Never imaged in any KMT2A-fusion cell. |
| cell type | Mouse embryonic stem cells and embryoid bodies; the human U937 monocytic line; 154 MDS bone marrows by qRT-PCR. No KMT2A-rearranged material of any kind. |
aliases retired or alternate
[ "IDAX" ]
domains
[
{
"name": "CXXC (CpG-binding zinc finger)",
"start": 132,
"end": 173,
"pfam": "PF02008"
}
]localization
{
"subcellular": "UNKNOWN for the human protein. UniProt's only term is Cytoplasm, ECO:0000250 (by similarity); there is no experimental localisation annotation, and the functional literature places IDAX at promoters and CpG islands, which contradicts it.",
"subnuclear": "not established",
"localization_shown_for_fusion": "NO"
}essentiality
{
"depmap_scan": "analyses/depmap-kmt2a-selective-scan-full",
"difference_kmt2a_r_minus_other": -0.0024,
"mean_kmt2a_rearranged": 0.0328,
"mean_other_leukemia": 0.0352,
"pan_essentiality_all_lines": 0.0093,
"n_all_lines": 1208,
"rank_of_17728": 8853,
"percentile": 49.938,
"expression_caveat": "CXXC4 is not expressed in ES cells and is induced on differentiation; a near-zero gene effect may mean the gene is not expressed in the screened lines rather than that it is dispensable",
"release_note": "Open Targets dataVersion 26.06, read 2026-09-21"
}- CXXC4 is 198 residues and is nearly all reader domain, which makes it the most likely family member to be hit hard by a CXXC-directed ligand and the least likely to tolerate it, since it has no other domain to work through.
- If CXXC4 delivers TET2 to DNA and then precedes its destruction, raising CXXC4 lowers TET2 protein. Any intervention that changes CXXC4 levels has a non-obvious sign on the methylome.
- Is human CXXC4 nuclear? UniProt says cytoplasm by similarity and the functional literature says CpG islands. One immunofluorescence experiment in a human haematopoietic cell would settle it and would also settle whether CXXC4 can be a competitor for the fusion's address at all.
- Is CXXC4 expressed in KMT2A-rearranged blasts? If not, it drops out of the competition question entirely, and that is answerable from existing infant ALL RNA-seq in an afternoon.
- Does the trans-encoded arrangement mean TET2's address can be modulated pharmacologically by changing CXXC4 dosage - a target no cis-domain protein offers?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
CXXC4 (Q9H2H0) is 198 aa with a CXXC-type zinc finger at 132-173; its curated DOMAIN comment states the CXXC mediates binding to CpG-DNA; and its only subcellular-location term is Cytoplasm with evidence code ECO:0000250, by similarity, with no experimental localisation annotation.
UniProtKB Q9H2H0 (CXXC4_HUMAN), CXXC-type zinc finger protein 4 / Inhibition of the Dvl and axin complex protein (IDAX) · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Fetched 2026-09-21; accession resolved by gene_exact:CXXC4 AND organism_id:9606 AND reviewed:true, one row, primary gene name read back as CXXC4 and length 198 from the same record. The absence of an experimental localisation is a statement about curation; a primary paper may contain one. No isoforms are annotated.
Evidence that argues against this (1)
- Ko et al. report IDAX localising to promoters and CpG islands in genomic DNA, which is a nuclear observation and directly contradicts the by-similarity cytoplasmic annotation. Ko M et al. 2013
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
IDAX/CXXC4 was originally encoded within an ancestral TET2 gene that underwent a chromosomal inversion separating the CXXC domain from the catalytic domain; the IDAX CXXC binds unmethylated CpG, localises to promoters and CpG islands, and interacts directly with the TET2 catalytic domain; IDAX expression causes caspase activation and TET2 protein down-regulation in a manner dependent on IDAX CXXC DNA binding.
Modulation of TET2 expression and 5-methylcytosine oxidation by the CXXC domain protein IDAX. · Ko M, An J, et al. · Nature 2013What this evidence cannot show
Mouse embryonic stem cells and the human monocytic line U937. The caspase-dependent TET2 down-regulation is an unusual mechanism reported by one group. The inversion is a synteny inference rather than an experiment. Nothing in leukaemia and nothing in a KMT2A-rearranged system.
Evidence that argues against this (1)
- In embryonic stem cells the paralogue Rinf/CXXC5 regulates Tet1 and Tet2 at the level of transcription by binding their promoters and enhancers, which is a different mechanism from escorting TET2 to DNA and then destabilising it. Ravichandran M et al. 2019
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Idax is not expressed in embryonic stem cells but is induced on differentiation, and Idax and Rinf together facilitate expression of Tet enzymes, promoting neural and suppressing trophectodermal programmes.
Idax and Rinf facilitate expression of Tet enzymes to promote neural and suppress trophectodermal programs during differentiation of embryonic stem cells. · Abou-Jaoude A, Huang CY, et al. · Stem Cell Res 2022What this evidence cannot show
Mouse embryonic stem cells and embryoid bodies; a developmental system with no haematopoietic or malignant component. Loss-of-function is by genetic deletion in ESCs; the readouts are lineage markers.
Evidence that argues against this (1)
- Ko et al. report that IDAX expression DOWN-regulates TET2 protein, while this study reports Idax facilitating Tet expression - the two are not contradictory only if transcript and protein move in opposite directions, which neither paper measures together. Ko M et al. 2013
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In 154 myelodysplastic-syndrome marrows against 50 controls, CXXC4 and TET2 mRNA were both significantly lower in patients, CXXC4 was higher in advanced WHO subtypes, and CXXC4 level was associated with overall survival but not with leukaemia-free survival and was not independent on multivariate analysis.
CXXC4 mRNA levels are associated with clinicopathological parameters and survival of myelodysplastic syndrome patients. · Li S, Fan R, et al. · Leuk Res 2014What this evidence cannot show
qRT-PCR on bulk marrow, so a difference could be a blast-fraction difference rather than a per-cell change. Single centre. The survival association does not survive multivariate adjustment, where only age, haemoglobin and marrow blast level remain, so this is a correlation and not a prognostic finding.
Evidence that argues against this (1)
- CXXC4 mRNA is reported LOWER than controls overall and HIGHER in advanced subtypes in the same study, which is internally awkward and consistent with bulk-composition effects. Li S et al. 2014
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
CXXC4 shows no dependency in KMT2A-rearranged cell lines: difference -0.0024, rank 8,853 of 17,728, pan-essentiality +0.0093.
Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap CRISPR Chronos via the Open Targets Platform GraphQL API) · Oncofusion knowledge-graph project · analyses/depmap-kmt2a-selective-scan-full/results.json, generated 2026-09-19; Open Targets dataVersion year 26 month 06, re-read from the API on 2026-09-21 2026What this evidence cannot show
The scan compares 11 Cellosaurus-annotated KMT2A-rearranged lines against 43 other leukaemia lines; none of the 11 is infant-derived and the arm is myeloid-enriched, so a difference mixes fusion status with lineage. Both the difference and the pan-essentiality across all 1,208 lines are quoted, because a gene required everywhere is not a selective dependency. Open Targets dataVersion 26.06; DepMap's own release identifier is not returned by that endpoint.
Evidence that argues against this (1)
- CXXC4 is not expressed in embryonic stem cells and is induced on differentiation, so a screen in established leukaemia lines may be run in cells that do not express it at all; DepMap gene effect is uninformative for a non-expressed gene. Abou-Jaoude A et al. 2022
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
The CXXC domains encoded at the CXXC4 and Cxxc10-1 loci, together with those of TET1 and CXXC5, form a distinct homology group within the CXXC family, and Tet3 interacts with Cxxc4.
Intrinsic and extrinsic connections of Tet3 dioxygenase with CXXC zinc finger modules. · Liu N, Wang M, et al. · PLoS One 2013What this evidence cannot show
Mouse; a sequence-homology grouping plus biochemical interaction. In-vitro binding on single-CpG substrates.
Evidence that argues against this (1)
- The systematic human family comparison classifies CXXC domains by measured DNA selectivity rather than by sequence homology, and those two classifications do not coincide. Xu C et al. 2018
- E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
The CXXC4 literature is small and the IDAX alias is partially contaminated: CXXC4[tiab] returns 51 records and IDAX[tiab] returns 16, of which the top-ranked titles include Mullerian-inhibiting-substance and computational-chemistry papers alongside the CXXC4 ones.
PubMed E-utilities esearch record counts with querytranslation inspected and every returned title read · Oncofusion knowledge-graph project · NCBI E-utilities, queries run 2026-09-21; exact query strings given in the caveat of each item that uses this source 2026What this evidence cannot show
Queries `CXXC4[tiab]` and `IDAX[tiab]`, run 2026-09-21; the first twelve IDAX titles were read individually. Title/abstract only, so per rule 80 a CXXC4 measurement inside a supporting figure is invisible. IDAX at n=16 is small enough to title-check exhaustively and that was done for the top twelve, not all sixteen.
Evidence that argues against this (1)
- A small record count for a gene symbol is not evidence of a small literature when the protein is usually discussed as part of the TET pathway; CXXC4 appears in TET2 papers that never put the symbol in the abstract. Ko M et al. 2013
- E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The string CXXC4 occurs zero times in 71,489 characters of the full text of the whole-genome study of infant KMT2A-rearranged ALL.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, et al. · Nat Genet 2015What this evidence cannot show
One paper, main text and reference list only; supplementary tables were not parsed. A zero in one study is not a zero in the disease.
Evidence that argues against this (1)
- That study reports a strikingly low mutation burden overall - a mean of 1.3 non-silent mutations in the dominant clone - so almost any gene would return zero and the absence carries less information than it appears to. Andersson AK et al. 2015
The IDAX CXXC domain interacts directly with the TET2 catalytic domain and supplies in trans the address module TET2 lacks in cis.
Binds unmethylated CpG and localises to promoters and CpG islands - although UniProt's only localisation term for the human protein is cytoplasm, by similarity.
One of the eleven other human CXXC domains a ligand for this interface must discriminate against, and the one that is almost nothing but the domain.
One of the twelve human CXXC proteins.
The two small trans-acting TET escorts; in differentiating ES cells Idax and Rinf act together on Tet expression.