EZH2 V662A Discovery
Novel unreported variant, SET domain loss-of-function, PRC2 sixth axis
Literature Hits Novel
0
0 for 'EZH2 V662A' or 'EZH2 V662' in Semantic Scholar; PubMed also 0 (rechecked 2 Nov 2026)
EVE Score Uncertain
0.7825
Evolutionary model, Frazer et al. Nature 2021. Label at EVE's class25 threshold.
AlphaMissense Likely pathogenic
0.9984
DeepMind proteome-wide missense predictor (its own class label)
ACMG Points Likely Pathogenic
9
Bayesian point system (Likely Pathogenic 6-9, Pathogenic 10+), 9 achieved; research classification, ClinVar: VUS
GENIE Carriers Absent
0 / 20,739
Not observed in the 20,739 GENIE v19.0 myeloid samples whose panel covers EZH2
Novel Discovery
EZH2 V662A has never been described in the scientific literature. A
search of PubMed and Semantic Scholar returns zero
results for "EZH2 V662A" or "EZH2 V662." Among the 314
EZH2 papers retrieved for this search, none report this specific
variant. ClinVar has a single record with no publications (below).
The variant is absent from gnomAD v4 (0 alleles across ~1.6 million alleles) and absent from GENIE v19.0 (0 carriers among the 20,739 myeloid samples whose panel covers EZH2, see the denominator table).
ClinVar holds one record for this variant: VCV003373649.1, submitted by GeneDx and classified as a variant of uncertain significance (germline, condition not provided), last evaluated March 2024. It carries no functional evidence, no disease context and no supporting publications. This project has not submitted the variant to ClinVar.
The variant has never been described in the indexed literature and has never been characterised in a somatic or myeloid context, in a gene that has been studied extensively for over 15 years in myeloid malignancies.
The variant is absent from gnomAD v4 (0 alleles across ~1.6 million alleles) and absent from GENIE v19.0 (0 carriers among the 20,739 myeloid samples whose panel covers EZH2, see the denominator table).
ClinVar holds one record for this variant: VCV003373649.1, submitted by GeneDx and classified as a variant of uncertain significance (germline, condition not provided), last evaluated March 2024. It carries no functional evidence, no disease context and no supporting publications. This project has not submitted the variant to ClinVar.
The variant has never been described in the indexed literature and has never been characterised in a somatic or myeloid context, in a gene that has been studied extensively for over 15 years in myeloid malignancies.
Computational Pathogenicity
| Model | Score | Classification | Reference |
|---|---|---|---|
| EVE | 0.7825 | Uncertain (at EVE's class25 threshold) | Frazer et al. Nature 2021 |
| AlphaMissense | 0.9984 | Likely pathogenic | Cheng et al. Science 2023 |
| CADD | 28.1 | Deleterious (threshold 25.3) | Schubach et al. Nucleic Acids Res 2024 (v1.7); threshold: Pejaver et al. Am J Hum Genet 2022 |
| REVEL | 0.781 | Pathogenic (threshold 0.644) | Ioannidis et al. Am J Hum Genet 2016; threshold: Pejaver et al. 2022 |
| ESM-2 LLR | -2.966 | PP3 Supporting | Lin et al. Science 2023 |
| SIFT | 0.0 | Deleterious (low confidence) | Ng & Henikoff 2003 |
| PolyPhen-2 | 0.992 | Probably damaging | Adzhubei et al. 2010 |
6 of 7 independent computational models classify EZH2 V662A as pathogenic or damaging, giving PP3_Strong. EVE does not: EVE returns 0.7825, which its own class25 calibration (25th/75th percentile ClinVar calibration (recommended)) labels Uncertain. CADD and REVEL are the Ensembl VEP CADD v1.7 values.
SET Domain Loss-of-Function
V662A (NP_004447.2 numbering; V657 in UniProt Q15910 numbering) falls
within the catalytic SET domain (residues 612 to 727, UniProt Q15910) of EZH2, the enzymatic subunit of the Polycomb
Repressive Complex 2 (PRC2). This domain is responsible for
trimethylation of histone H3 at lysine 27 (H3K27me3), a critical
epigenetic silencing mark.
Chase et al. 10
Ernst et al. 9
Chase et al. 10
[10] A 2020
Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020)
and the same group's earlier in vitro
work showed loss or greatly reduced H3K27 methylation for the SET/CXC
missense mutants they tested (C576W, R690C, Y731D). V662 itself was not
tested, so loss of function for V662A is inferred from its position in
the SET domain, not measured. This is the opposite mechanism from lymphoid
malignancies (DLBCL), where EZH2 gain-of-function mutations at Y646
increase H3K27me3. The tissue-specific directionality is critical for
therapeutic decisions.
Ernst et al. 9
[9] T 2010
Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet (2010)
first described EZH2 inactivating
mutations in myeloid disorders (2010, 1,195 citations). The myeloid
context is now well established: EZH2 acts as a tumor suppressor, and
loss-of-function mutations contribute to disease through derepression
of Polycomb target genes.
Tazemetostat: CONTRAINDICATED
Tazemetostat (Tazverik) is an EZH2 inhibitor that held FDA accelerated
approval in relapsed or refractory follicular lymphoma (EZH2-mutant, or
any EZH2 status without other options) and in epithelioid sarcoma, a
tumour driven by SMARCB1/INI1 loss, not EZH2 mutation. It works by
inhibiting EZH2 enzymatic activity.
EZH2 V662A is inferred to be a loss-of-function mutation. Tazemetostat inhibits the very activity that is already lost. Administering it would further reduce residual H3K27me3, potentially worsening the epigenetic dysregulation.
This is not a theoretical concern. Ipsen voluntarily withdrew tazemetostat in all indications from all its markets on March 9, 2026, after hematologic second cancers (mostly MDS and AML) emerged in the SYMPHONY-1 trial, and the distinction between gain-of-function (lymphoid) and loss-of-function (myeloid) EZH2 mutations is fundamental to the biology 10
Clinical implication: Any clinician seeing "EZH2 mutation" in a myeloid patient must verify the functional consequence before considering EZH2-targeted therapy. The mutation name alone is insufficient; the mechanism (GoF vs LoF) determines whether therapy helps or harms.
EZH2 V662A is inferred to be a loss-of-function mutation. Tazemetostat inhibits the very activity that is already lost. Administering it would further reduce residual H3K27me3, potentially worsening the epigenetic dysregulation.
This is not a theoretical concern. Ipsen voluntarily withdrew tazemetostat in all indications from all its markets on March 9, 2026, after hematologic second cancers (mostly MDS and AML) emerged in the SYMPHONY-1 trial, and the distinction between gain-of-function (lymphoid) and loss-of-function (myeloid) EZH2 mutations is fundamental to the biology 10
[10] A 2020
Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020)
.
Clinical implication: Any clinician seeing "EZH2 mutation" in a myeloid patient must verify the functional consequence before considering EZH2-targeted therapy. The mutation name alone is insufficient; the mechanism (GoF vs LoF) determines whether therapy helps or harms.
Research classification: Likely Pathogenic (ClinVar: VUS)
EZH2 V662A was initially classified as a Variant of Uncertain
Significance (VUS) due to the absence of direct functional assay data
(PS3 at Supporting level only, no definitive functional study of V662A
specifically).
Re-running the ACMG aggregation against the current evidence files gives Likely Pathogenic at 9 Bayesian points (Likely Pathogenic from 6, Pathogenic from 10). This is a research classification; the public ClinVar record remains a VUS. The evidence lines it draws on, and the points each one contributes, are listed in the table below.
The computational evidence line is PP3_Strong, on 6/7 concordant predictors: AlphaMissense 0.9984, CADD 28.1, REVEL 0.781, ESM-2 LLR -2.966, SIFT and PolyPhen-2. EVE is the dissenting tool at 0.7825, which its own class25 calibration labels Uncertain. In-silico evidence is capped at Strong (ClinGen SVI, Pejaver et al. 2022), so the computational line adds 4 points and does not depend on the EVE vote.
With this classification the patient profile is 4 confirmed drivers plus EZH2 V662A as Likely Pathogenic (research classification, without a functional assay of V662A itself).
Re-running the ACMG aggregation against the current evidence files gives Likely Pathogenic at 9 Bayesian points (Likely Pathogenic from 6, Pathogenic from 10). This is a research classification; the public ClinVar record remains a VUS. The evidence lines it draws on, and the points each one contributes, are listed in the table below.
The computational evidence line is PP3_Strong, on 6/7 concordant predictors: AlphaMissense 0.9984, CADD 28.1, REVEL 0.781, ESM-2 LLR -2.966, SIFT and PolyPhen-2. EVE is the dissenting tool at 0.7825, which its own class25 calibration labels Uncertain. In-silico evidence is capped at Strong (ClinGen SVI, Pejaver et al. 2022), so the computational line adds 4 points and does not depend on the EVE vote.
With this classification the patient profile is 4 confirmed drivers plus EZH2 V662A as Likely Pathogenic (research classification, without a functional assay of V662A itself).
| ACMG Criterion | Strength | Points | Evidence |
|---|---|---|---|
| PP3 | Strong | 4 | 6/7 pathogenic: AlphaMissense=0.9984 (PP3_Strong), CADD=28.1, REVEL=0.781, ESM-2 LLR=-2.966, EVE=0.7825 (Uncertain), SIFT=0.0, PolyPhen-2=0.992 [capped at Strong: ClinGen SVI, Pejaver et al. 2022, Am J Hum Genet 109:2163-2177, PMID 36413997] |
| PM1 | Moderate | 2 | Catalytic SET domain (612-727). Critical for H3K27 methyltransferase activity. |
| PS3 | Supporting | 1 | V662A in catalytic SET domain (612-727). Loss-of-function reducing H3K27me3. EZH2 LOF established as tumor suppressor in myeloid lineage. Ernst 2010, Nikoloski 2010. No direct functional assay for V662A specifically (PS3 at Supporting level). |
| PM2 | Supporting | 1 | Absent from gnomAD v4 (AC=0). Completely absent from gnomAD (0 alleles across exomes + genomes) |
| PM5 | Supporting | 1 | Other missense variants in the EZH2 SET domain (e.g., Y646, A682, A692) are established as pathogenic in lymphoid malignancies (gain-of-function). In myeloid context, EZH2 LOF missense across SET domain are recurrent. |
| Total | 9 | Likely Pathogenic (threshold: 10) |
ACMG/AMP Bayesian point system (Tavtigian et al. Hum Mutat 2020). Richards et al. Genet Med 2015. Rows are read from acmg_aggregation.json as regenerated on 2026-11-03; no value on this page is entered by hand.
PRC2: The 6th Convergence Axis
With EZH2 V662A classified Likely Pathogenic, the patient's mutation profile
spans 6 biological convergence axes, not 5:
1. DNA methylation (DNMT3A R882H)
2. Metabolic (IDH2 R140Q, 2-HG production)
3. PP2A tumor suppression (SETBP1 G870S)
4. RAS-MAPK signaling (PTPN11 E76Q)
5. Cytogenetic (monosomy 7)
6. Chromatin remodeling / PRC2 (EZH2 V662A)
The PRC2 axis is mechanistically distinct from the DNA methylation axis (DNMT3A) and the metabolic axis (IDH2). While all three affect the epigenetic landscape, they operate through independent biochemical pathways: CpG methylation, histone H3K27 trimethylation, and alpha-ketoglutarate-dependent dioxygenase inhibition, respectively.
This 6-axis convergence on a single patient is, to our knowledge, unprecedented in the myeloid genomics literature.
1. DNA methylation (DNMT3A R882H)
2. Metabolic (IDH2 R140Q, 2-HG production)
3. PP2A tumor suppression (SETBP1 G870S)
4. RAS-MAPK signaling (PTPN11 E76Q)
5. Cytogenetic (monosomy 7)
6. Chromatin remodeling / PRC2 (EZH2 V662A)
The PRC2 axis is mechanistically distinct from the DNA methylation axis (DNMT3A) and the metabolic axis (IDH2). While all three affect the epigenetic landscape, they operate through independent biochemical pathways: CpG methylation, histone H3K27 trimethylation, and alpha-ketoglutarate-dependent dioxygenase inhibition, respectively.
This 6-axis convergence on a single patient is, to our knowledge, unprecedented in the myeloid genomics literature.
EZH2 Literature Landscape
| Query | Hits |
|---|---|
| EZH2 V662A | 0 |
| EZH2 V662 | 0 |
| EZH2 loss-of-function myeloid | 100 (capped) |
| EZH2 mutation MDS AML | 100 (capped) |
| PRC2 myeloid malignancy | 77 |
| EZH2 monosomy 7 | 61 |
| EZH2 SET domain missense | 22 |
Semantic Scholar query results, retrieval capped at 100 papers per query, so "(capped)" means at least 100. 314 unique papers retrieved across all EZH2-related queries.
The EZH2 literature is extensive (this capped search alone retrieved
314 unique papers), with landmark studies including Ernst et al.
9
Ball et al. 11
[9] T 2010
Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet (2010)
(1,195 citations, first description
of EZH2 inactivation in myeloid disorders) and Chase et al.
10[10] A 2020
Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020)
(SET domain LoF mechanism).
Ball et al. 11
[11] S 2023
Clinical characteristics and outcomes of EZH2-mutant myelodysplastic syndrome: a large single institution analysis of 1774 patients. Leuk Res (2023)
showed that, within EZH2-mutant MDS,
concurrent chromosome 7 abnormalities were associated with significantly
worse overall survival (median 20.8 vs 35.5 months, p=0.002), the
shortest median reported in that abstract. This patient has both: EZH2
V662A on monosomy 7, placing them in that subgroup.
Clinical Implications
Classifying EZH2 V662A as Likely Pathogenic has three immediate
clinical consequences:
1. Tazemetostat contraindication. Any EZH2-targeted therapy designed to inhibit EZH2 activity is contraindicated. The enzymatic activity is already lost.
2. Prognostic impact. EZH2 LoF mutations with chromosome 7 abnormalities carry inferior survival 11
3. Profile rarity. With EZH2 V662A counted as the fifth variant, 0 of 8,655 GENIE patients carry the profile, which bounds its frequency below 1 in 2,889 (95%, exact binomial); no point estimate exists. The profile was already unprecedented as a quadruple; as a quintuple it is genuinely unknown territory for prognostic modeling.
The classification rests on in silico evidence (PP3_Strong) rather than direct functional assay data. No ClinVar submission has been made; the public record for this variant remains VCV003373649.1, GeneDx's germline VUS of March 2024.
1. Tazemetostat contraindication. Any EZH2-targeted therapy designed to inhibit EZH2 activity is contraindicated. The enzymatic activity is already lost.
2. Prognostic impact. EZH2 LoF mutations with chromosome 7 abnormalities carry inferior survival 11
[11] S 2023
Clinical characteristics and outcomes of EZH2-mutant myelodysplastic syndrome: a large single institution analysis of 1774 patients. Leuk Res (2023)
. In that cohort this subgroup had the
shortest reported median survival among EZH2-mutant MDS (20.8 months).
3. Profile rarity. With EZH2 V662A counted as the fifth variant, 0 of 8,655 GENIE patients carry the profile, which bounds its frequency below 1 in 2,889 (95%, exact binomial); no point estimate exists. The profile was already unprecedented as a quadruple; as a quintuple it is genuinely unknown territory for prognostic modeling.
The classification rests on in silico evidence (PP3_Strong) rather than direct functional assay data. No ClinVar submission has been made; the public record for this variant remains VCV003373649.1, GeneDx's germline VUS of March 2024.
References
- Ernst T, Chase AJ, Score J, et al. Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet (2010). PubMed
- Chase A, Score J, Lin F, et al. Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020). PubMed
- Ball S, Aguirre LE, Jain AG, et al. Clinical characteristics and outcomes of EZH2-mutant myelodysplastic syndrome. Leuk Res (2023). PubMed
- Frazer J, Notin P, Dias M, et al. Disease variant prediction with deep generative models of evolutionary data. Nature (2021). PubMed
- Cheng J, Novati G, Pan J, et al. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science (2023). PubMed
- Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants. Genet Med (2015). PubMed
- Tavtigian SV, Harrison SM, Boucher KM, Biesecker LG. Fitting a naturally scaled point system to the ACMG/AMP variant classification guidelines. Hum Mutat (2020). PubMed